Received: 10 December 2018
Accepted: 22 March 2019
Is rTMS effective for anxiety symptoms in major depressive
disorder? An efficacy analysis comparing left‐sided
high‐frequency, right‐sided low‐frequency, and sequential
bilateral rTMS protocols
Kate E. Hoy
Paul B. Fitzgerald
Monash Alfred Psychiatry Research Centre,
Monash University, Melbourne, Victoria,
Epworth Centre for Innovation in Mental
Health, Epworth HealthCare, Camberwell,
Alfred Mental and Addiction Health, Alfred
Health, Melbourne, Victoria, Australia
Leo Chen, Monash Alfred Psychiatry Research
Centre, Monash University, Melbourne, VIC
National Health and Medical Research
Council, Grant/Award Numbers:
Background: Anxiety symptoms are common in major depressive disorder. Whilst
therapeutic efficacy of repetitive transcranial magnetic stimulation (rTMS) in
depression is well‐established, minimal research has investigated rTMS's efficacy in
treating anxiety symptoms in depression.
Methods: This study investigates the effectiveness of rTMS in treating anxiety
symptoms in depression, specifically the relative efficacy of the three rTMS protocols
commonly used in clinical practice: left‐sided high‐frequency, right‐sided low‐
frequency and sequential bilateral rTMS. Antidepressant efficacy of each rTMS
protocol is also investigated. Treatment data for 697 patients were pooled from three
studies across five sites. Changes in Beck's Anxiety Inventory (BAI) and the Hamilton
Depression Rating Scale over 4‐week rTMS courses were analysed using latent
growth curve modelling.
Results: All rTMS protocols were effective in treating anxiety symptoms (mean BAI
reduction, 8.13 points; p< 0.001) and depressive symptoms. Near therapeutic
equivalence was seen across the three protocols. Improvement in depressive severity
positively correlated with improvement in anxiety. Both high‐and low‐baseline
anxiety scores showed overall symptom reduction.
Conclusions: This study addresses the clinical knowledge gap pertaining to rTMS's
therapeutic efficacy in treating anxiety symptoms in depression and the relative
efficacy of three commonly used stimulation protocols. Our findings suggest
therapeutic equivalence across left‐sided high‐frequency, right‐sided low‐frequency,
and sequential bilateral rTMS approaches.
anxiety, anxiety disorders, brain stimulation, depression, repetitive transcranial magnetic
stimulation, treatment‐resistant depression
Depression and anxiety are common worldwide and associated with
significant disease burden, as assessed by disability‐adjusted life
years (Whiteford et al., 2013) and years of life lived with disability
(A. J. Baxter, Vos, Scott, Ferrari, & Whiteford, 2014). Persistent
illness is associated with significant subjective distress, functional
decline, increased morbidity and mortality as well as carer burden.
Depress Anxiety. 2019;1–9. wileyonlinelibrary.com/journal/da © 2019 Wiley Periodicals, Inc.
Despite their prevalence and negative impact on function and quality
of life, both can be refractory to conventional treatments, with
treatment resistance reported at rates of 30% for depressive (Crown
et al., 2002) and 25% for anxiety disorders (Kessler et al., 2005). The
two conditions often coexist, either as concomitant disorders or a
primary depressive or anxiety syndrome that manifest in secondary
depressive and anxiety symptoms. Lifetime comorbidity of the two
syndromes has been reported at 90% (Gorman, 1996). Neuroimaging
points to several shared brain regions that exhibit abnormal activity
across the two disorders, with one instance being overactivation of
the amygdala in patients presenting with either depressive symptoms
or anxiety responses (Ressler & Mayberg, 2007). Brain regions
implicated in mood regulation, such as the anterior cingulate and
prefrontal cortices (PFC), amygdala, insula, and limbic system have
also been implicated in anxious responses such as monitoring,
vigilance, and the expression and modulation of fear responses
(Etkin, 2012). The same neuroanatomical regions have also been
implicated in anxiety disorders such as generalised anxiety disorder
(GAD), obsessive‐compulsive disorder (OCD), posttraumatic stress
disorder (PTSD; Pallanti & Bernardi, 2009), and depressive disorders
(L. R. Baxter et al., 1989; George, Ketter, & Post, 1993, 1994).
Given the considerable disease burden and treatment‐refractory
nature of these disorders, clearly, the need exists for effective
treatment options in addition to what is available. One such modality
is repetitive transcranial magnetic stimulation (rTMS), a technique that
applies patterned electromagnetic pulses to superficial brain regions to
depolarise underlying neurons in a noninvasive manner. Repeated
stimulation modulates neuronal activity and downregulates electro-
chemical properties along neuronal pathways, which is one of the
postulated mechanisms of action behind its therapeutic potential in
depressive disorders (Liston et al., 2014; Tik et al., 2017). In clinical
practice, rTMS is typically applied to the left and/or right dorsolateral
aspects of the prefrontal cortex (DLPFC), inducing subtle, but
repeatable, changes in electrophysiological, metabolic, and intrasynaptic
activity both locally and distally along brain circuits implicated in mood
regulation (George et al., 2003). Indeed, as an antidepressant therapy,
rTMS has been extensively investigated over the past two decades, with
substantive evidence supporting its efficacy derived from multiple
sham‐controlledrandomisedtrials(Fitzgerald et al., 2003; George et al.,
2010; O'Reardon, Solvason, & Janicak, 2007) and meta‐analyses (Berlim,
van den Eynde, Tovar‐Perdomo, & Daskalakis, 2014; Schutter, 2009;
Slotema, Blom, Hoek, & Sommer, 2010). The three evidence‐based
stimulation parameters/targets commonly used in clinical settings are
high‐frequency stimulation to the left DLPFC (LHF rTMS), low‐
frequency stimulation to the right DLPFC (RLF rTMS), and bilateral
rTMS sequentially applied to both right and left DLPFC (BL rTMS), with
established equivalence of antidepressant efficacy across the three
protocols (J. Chen et al., 2013; J.‐J. Chen et al., 2014; Fitzgerald et al.,
2011, 2013; Hoy, Segrave, Daskalakis, & Fitzgerald, 2012). In contrast,
there has been a paucity of studies examining rTMS's effects on anxiety,
either as comorbid symptoms in major depressive disorder (MDD) or in
primary anxiety disorders such as GAD, OCD, and PTSD, but to name a
few. Four recent systematic reviews of open‐label and sham‐controlled
studies of rTMS and its effects on OCD, PTSD, GAD, panic disorder, and
social anxiety disorder failed to draw definitive conclusions on its clinical
efficacy (Machado et al., 2012; Pallanti & Bernardi, 2009; Pigot, Loo, &
Sachev, 2008; Zwanzger, Fallgatter, Zavorotnyy, & Padberg, 2009).
Some reported reasons for this are (a) the diverse variations of rTMS
protocols studied, (b) small sample sizes ranging between 10 and 42,
particularly in the case of sham‐controlled trials, and (c) the overall
limited number of studies.
Similarly, limited studies have examined the effects of rTMS on
comorbid anxiety symptoms that occur in depression, sometimes
referred to as “anxious depression.”In one study, a small subset of
patients with anxious depression was found to derive equal
antidepressant effects from LHF rTMS as nonanxious patients with
depression (Diefenbach, Bragdon, & Goethe, 2013). In another open‐
label study, 15 chronically depressed, severely treatment‐resistant
patients were treated with 20 sessions of LHF rTMS (Berlim, McGirr,
Beaulieu, & Turecki, 2011), resulting in 25% and 18.3% reductions
across two anxiety symptomatology rating scales. A third open‐label
trial of 36 depressed patients reported 15 sessions of left‐sided
20 Hz rTMS resulted in large therapeutic effect sizes (Pearson's
r> 0.5) on both anxiety and depression symptoms (Durmaz, Ebrinc,
Ates, & Algul, 2017). Contrary to these positive findings, however,
there have also been reports of worsening anxiety associated with
rTMS treatment (Greenberg, McCann, Benjamin, & Murphy, 1997).
The only systematic review and meta‐analysis specifically addressing
rTMS's effects on comorbid anxiety symptoms in depression was
published by coinventors of a proprietary TMS coil capable of deeper
cortical stimulation compared with standard figure‐of‐eight coils
(Kedzior, Gellersen, Roth, & Zangen, 2015). In this study, deep rTMS
was reported to have a large positive effect size (Cohen's dof 1.45) in
treating anxiety symptoms in treatment‐resistant MDD.
In sum, the small sample sizes, open‐label design, and hetero-
geneous rTMS parameters featured in these studies leave the
question of whether rTMS is effective in treating anxiety symptoms
in depression unanswered. It is also not known which rTMS protocol
(s) are superior in this regard. Given the ubiquitous occurrence of
anxiety symptoms in depression sufferers, evidence to guide optimal
rTMS protocol choice is timely and can be of significant clinical value.
Therefore, we aimed to investigate the efficacy of rTMS in treating
comorbid anxiety symptoms in MDD over a 4‐week treatment
course. A three‐way comparison of relative efficacy over time across
LHF rTMS, RLF rTMS, and BL rTMS protocols was also investigated.
We analysed pooled data from three clinical trials that featured a
combination of RLF/BL rTMS (Fitzgerald et al., 2011), RLF/BL rTMS
(Fitzgerald et al., 2013), and LHF/RLF rTMS (unpublished data)
protocols. Data from the last of these trials were recently submitted
for publication and currently under review. The aim of this study is to
determine if rTMS is effective in treating anxiety symptoms in
CHEN ET AL.
depression and if so, if one rTMS protocol is superior to others. To
this end, pooled data from multiple trials of comparable patient
demographics, diagnoses, symptom severity, rTMS protocols, and
identical outcome measures offer a larger data set for statistical
analysis. All three studies utilised the 17‐item Hamilton Depression
Rating Scale (HAMD) to measure change in depression symptoma-
tology and the 21‐item Beck's Anxiety Inventory (BAI) to measure
change in anxiety symptoms. All trials were randomised, double‐blind
studies of prospective design, where patients and raters were blind
to the treatment administered, but not the clinicians providing
treatment. Randomisation was achieved using a single computer‐
generated random number sequence (no stratification). Patients and
raters were advised there was a difference in the stimulation
parameters across the randomisation groups but specific information
related to stimulation sites, sides and durations were not described in
detail. Across all rTMS protocols, patients received courses of 20
treatment sessions, 5‐days‐a‐week over 4 weeks.
Treatment data for 697 patients were pooled for analysis from
three trials, of individual sample sizes of 179, 218, and
300 patients. Treatment outcomes were pooled then separated
into three treatment groups: LHF rTMS, RLF rTMS, and BL rTMS.
Patient age, gender distribution, and psychiatric diagnoses were
comparable across the three treatment groups (Table 1).
Psychiatric diagnoses were determined using the Mini‐Interna-
tional Neuropsychiatric Interview (Sheehan et al., 1998) by a study
psychiatrist for each patient. Patients were categorised as having:
MDD, single episode (n= 170); MDD, relapse (n= 415); bipolar I
disorder (BPAD I), depressive episode (n= 50); bipolar II disorder
(BPAD II), depressive episode (n= 34). Only patients with depres-
sion severity in the moderate‐severe range (scoring >13 on the
HAMD; Bech, Kastrup, & Rafaelsen, 1986) were included. All
patients had failed to respond to at least two antidepressant
medications for at least 6 weeks in the current illness episode,
qualifying their illness as stage II treatment‐resistant depression
(Thase & Rush, 1997). Exclusion criteria included significant active
medical illness, significant comorbid substance use disorder,
current neurological disease, or a contraindication to undergo
rTMS for safety reasons, such as a history of seizures and presence
of metal anywhere in the head/facial regions. Past failure to
respond to electroconvulsive therapy was not an exclusion
criterion. Psychotropic medications were not allowed to have
changed in the 4 weeks before the commencement of the trial or
during the trial itself.
Across the three trials, patients were recruited by referral
from psychiatrists and other medical practitioners between
January 2006 and October 2015. All three trials were conducted
in inpatient settings across five private psychiatric hospitals in
the three Australian states of Victoria, New South Wales, and
Queensland. Training in rTMS methods, trial management, and
data collection were conducted by one lead site to assure
uniformity of delivery across the sites. After complete descrip-
tion of the original studies to the patients, written informed
consent was obtained from all patients and each study received
Human Ethics Committee approval.
Across all study sites, rTMS was administered using MagVenture
MagPro R30 magnetic stimulators (MagVenture Inc., Farum, Den-
mark) using fluid‐filled 70‐mm figure‐of‐8 coils held by stands. The
coils were held tangential to the scalp at 45° to the midline. The
location for stimulation was a point 6 cm anterior to that required for
maximum stimulation of the Abductor Pollicis Brevis muscle. The
resting motor threshold (RMT) was measured using standard visual
methods (Pridmore, Fernandes Filho, Nahas, Liberatos, & George,
1998). Patients sat in a comfortable reclining chair during all
The TMS treatment protocols are presented in Table 2. Across
the three pooled studies, 150 patients received LHF rTMS, 312
received RLF rTMS, and 235 received BL rTMS.
The primary outcome measure for all three studies was the
HAMD score. The BAI was a secondary outcome measure across
the studies. Assessment scores were obtained at baseline and
TABLE 1 Participant demographic and clinical characteristics
Age in years,
46.2 (12.7) 44.9 (14.5) 48.7 (15.0)
Gender, N(F/M) 105/45 214/98 166/69
Gender (%F/%M) 70.0/30.0 68.6/31.4 70.6/29.4
48 70 52
MDD relapse 79 199 137
BPAD I 6 23 21
BPAD II 4 9 21
131 (87.3) 263 (84.3) 195 (83.0)
HAMD, mean (SD) 26.6 (6.4) 23.0 (6.0) 21.0 (5.3)
BAI, mean (SD) 30.6 (13.7) 28.3 (13.2) 25.8 (12.3)
Note. BAI: 21‐item Beck's Anxiety Inventory; BPAD I: bipolar I disorder;
BPAD II: bipolar II disorder; HAMD: 17‐item Hamilton Depression Rating
Scale; MDD: major depressive disorder; rTMS: repetitive transcranial
magnetic stimulation; SD: standard deviation.
CHEN ET AL.
after 2 and 4 weeks. As such, these time points were used in the
This analysis addresses differences in gender‐adjusted effects of LHF
rTMS, RLF rTMS, and BL rTMS on anxiety‐and depression‐severity
scores over time. We conducted a conditional two‐part latent growth
model (LGM) analysis, modelling structural equations of BAI and
HAMD scores at three time points of each (baseline, Weeks 2 and 4).
This analysis strategy differs from repeated measure analyses of
variance (ANOVA) because here, we model longitudinal manifest
outcome variables with latent or unobserved random intercepts and
slopes. Fundamentally different from ANOVA, LGM accounts for
each subject's individual score trajectory and allow for associative
equation building where conditional time‐variable scores of both BAI
and HAMD are fitted in one model (Figure 1). Furthermore, errors
(e) or disturbances (D) variances were freely parameterised.
We first fitted a random intercept‐only (or base model) where
scores of both scales assumed no‐growth. Second, a linear growth
trajectory was modelled with fixed and equally spaced factor
loadings (λslope = 0, 1, and 2). Finally, we tested a freed‐loading or
nonlinear model (λslope = 0, 1, and *). Incremental (Comparative Fit
Index [CFI] >0.90 and Tucker–Lewis Index [TLI] >0.90) and absolute
fit indices (Akaike's Information Criteria [AIC], χ
“smaller the better “across models, and root mean square error of
approximation [RMSEA], <0.09) were utilised to determine the best‐
fitting model (Curran & Muthen, 1999). Analyses used full informa-
tion maximum likelihood estimator method, missing data was
assumed to be missing at random. For hypothesis testing, the two‐
tailed αlevel was set at 0.05. Models were written with CALIS
procedure, SAS (SAS institute, Cary, NC).
There were no significant differences in patient demographics and
clinical variables across the three studies or three rTMS protocols.
Differences in baseline HAMD and BAI scores were not clinically
significant across the three rTMS protocols. Specifically, the mean
difference ranges were 5.6 points for HAMD and 4.8 for BAI
(Table 1) across the groups. Nonetheless, to increase model precision,
we adjusted for baseline difference by fitting a subject‐specific
trajectory model. Further, in the statistical modelling, we imposed a
covariance matrix that allows baseline latent parameters (intercepts)
to covary, adding to the robustness of the treatment estimates.
The strategy for statistical model selection was incremental. First, we
used the outfit option of the data step to estimate the base model
with random intercept‐only (CFI = 0.00). The base‐fit option written
with subsequent models affirmed the selection of freed loadings or
nonlinear model as a stable solution with excellent incremental
(relative) and absolute fit indices: CFI = 0.98, TLI = 0.95, and
RMSEA = 0.06 (Table 3).
Table 4 gives the parameter estimates for the model including
gender as a covariate. Of interest, female patients are associated
with higher baseline BAI and HAMD scores (β= 0.19 and 0.14,
respectively). However, the projection of BAI and HAMD slopes over
the courses of rTMS is independent of patient gender. There is
statistically significant evidence that LHF rTMS resulted in greater
reduction of HAMD growth scores relative to RLF rTMS (β=−0.25;
p= 0.009) and BL rTMS (β=−0.35; p< 0.001), after controlling for
gender. In comparison, the three rTMS protocols exhibited minimal
clinically significant relative difference in reducing BAI growth scores
TABLE 2 rTMS protocols across the three pooled studies
Study 1: Fitzgerald
et al. (2011)
Study 2: Fitzgerald et al.
Study 3: Fitzgerald et al.
Sample size N= 218 N= 179 N= 300
rTMS Protocol RLF rTMS (n= 71) R priming then LF rTMS (n= 91) LHF rTMS 50 trains (n= 59)
BL rTMS (n= 71) BL rTMS (n= 88) LHF rTMS 125 trains (n= 91)
BL 1 Hz rTMS (n= 76) –RLF rTMS 20 min (n= 57)
RLF rTMS 60 min (n= 93)
Localisation method “6 cm rule”“6 cm rule”“6 cm rule”
Stimulation hemisphere and frequency 1 Hz right 6 Hz priming right then 1 Hz right 10 Hz left
1 Hz right then 10 Hz left 1 Hz right then 10 Hz left 10 Hz left
1 Hz right then 1 Hz left –1 Hz right
1 Hz right
Stimulation intensity (%RMT) 110 90 for priming stimulation 110 120
Number of pulses per stimulation session RLF rTMS = 900 R priming then LF rTMS = 1,500 LHF rTMS 50 trains = 2,250
BL rTMS = 1,800 BL rTMS = 1,650 LHF rTMS 125 trains = 5,625
BL rTMS = 1,800 –RLF rTMS 20 min = 1,200
RLF rTMS 60 min = 3,600
Note. BL rTMS: sequential bilateral repetitive transcranial magnetic stimulation; LHF rTMS: left‐sided high‐frequency repetitive transcranial magnetic
stimulation; RLF rTMS: right‐sided low‐frequency repetitive transcranial magnetic stimulation; RMT: resting motor threshold; rTMS: repetitive
transcranial magnetic stimulation.
CHEN ET AL.
(βin a range of −0.08 to 0.07), which were also not statistically
Covariances indicate patients with higher HAMD scores at
baseline also scored higher BAI at baseline (p< 0.001). Likewise,
change in HAMD scores over time positively correlated with change
in BAI (p< 0.001). These findings suggest a high degree of
concordance between HAMD and BAI scores both at baseline and
in their treatment response trajectories. Both high‐and low‐baseline
BAI showed overall symptom reduction as measured by change in
BAI and HAMD ( Figures 1 and 2).
Change in anxiety (BAI) over time across
three rTMS protocols (Figure 2)
All three rTMS protocols resulted in a reduction in anxiety severity,
as measured by BAI, with a mean reduction of 8.13 points (p< 0.001).
Relative to RLF rTMS and BL rTMS, LHF rTMS appeared to
demonstrate slightly increased anxiolytic effects (b=−0.38 and
−1.50, respectively, but these were not clinically or statistically
significant. BL rTMS appeared to be marginally inferior to RLF rTMS
in antianxiolytic effects, although again, this was not clinically or
FIGURE 1 Conditional two‐part latent growth curve model (LGM) describing associations between anxiety (BAI) and depression (HAMD)
scores across three time points (baseline [B], 2 weeks [W2], 4 weeks [W4]). We used the structural models over ANOVA, mixed and marginally
analysed as the former accounted for error variances and score heterogeneity in a large data set. Individual‐specific growth curves and response
variables are also captured by LGM. Further, missing data are treated as missing at random (MAR), whereas ANOVA assumes missing
completely at random (MCAR) and performs a list‐wise deletion. In our freed‐loading model, factor loadings on each time point were all above
0.80 (p< 0.001). Error variances were freely estimated. Covariances between BAI and HAMD latent structures (slopes and intercepts) are
shown above with coefficients. BAI and HAMD growth processes were conditioned on three‐way rTMS variables and gender (dummy coded
with Δ, dotted lines). ANOVA: analysis of variance; BAI: 21‐item Beck's Anxiety Inventory; HAMD: 17‐item Hamilton Depression Rating Scale;
rTMS: repetitive transcranial magnetic stimulation
TABLE 3 Model fit information (N= 697)
Number of parameters DF AIC CFI TLI RMSEA (90% CI)
Base 1,507.56* 26 28 27694.29 0.00 −0.38 0.27 (0.26–0.29)
Linear 196.11* 41 13 26412.84 0.87 0.73 0.14 (0.12–0.16)
Nonlinear 42.86* 43 11 26263.59 0.98 0.95 0.06 (0.04–0.08)
Note. Base is the random intercept‐only model, linear is the fixed loading model, and nonlinear is the freed‐loading model.
AIC: Akaike's Information Criterion; CFI: Comparative Fit Index; CI: confidence interval; DF: χ
degrees of freedom; RMSEA: root mean square error of
approximation; rTMS: repetitive transcranial magnetic stimulation; TLI: Tucker–Lewis Index.
CHEN ET AL.
statistically significant. Gender was not a statistically significant
determining factor in anxiety reduction.
Change in depression (HAMD) over time
across three rTMS protocols (Figure 3)
All three rTMS protocols resulted in antidepressant effects with a
mean reduction of 8.61 points on the HAMD, which was statistically
significant (p< 0.001). Relative to RLF rTMS, LHF rTMS reduced
HAMD by a further 2.20 mean points (p= 0.009), and a further 3.05
mean points (p< 0.001) relative to BL rTMS. BL rTMS demonstrated
marginally reduced antidepressant efficacy, relative to RLF rTMS,
although this was not statistically significant (p= 0.48). Gender was not
a statistically significant determining factor in antidepressant response.
To our knowledge, this is the largest study to date that specifically
investigated rTMS's efficacy in treating anxiety symptoms in MDD
and if anxiolytic efficacy differs across the protocols commonly used
in clinical practice. Our results suggest that RLF rTMS, LHF rTMS,
and BL rTMS are all equally effective. Our findings are in keeping
with the earlier, smaller studies that have found rTMS to be effective
in treating comorbid anxiety occurring in MDD (Berlim, McGirr, et al.,
2011; Diefenbach et al., 2013; Durmaz et al., 2017; Greenberg et al.,
1997; Kedzior et al., 2015). With respect to treatment efficacy of
depressive symptoms, our results slightly favour LHF rTMS over BL
rTMS and RLF rTMS, which is not necessarily consistent with
previous reports (J. Chen et al., 2013; J.‐J. Chen et al., 2014;
Fitzgerald et al., 2011, 2013; Hoy et al., 2012). Female gender was
not a statistically or clinically significant predictor of response for
both anxiety and depression symptoms. Adding to the generalisa-
bility of our results is the fact that our treatment outcomes were
derived from five private psychiatric hospitals across three Austra-
lian states over an almost 10‐year period.
In contrast to evidence supporting its antidepressant efficacy,
rTMS's therapeutic potential in anxiety disorders can be considered
equivocal and promising at best (Pallanti & Bernardi, 2009), despite
neuroimaging evidence of shared regions of aberrant brain activity
and connectivity in the expression of depressive and anxiety
symptoms. It is also curious that certain antidepressant classes (such
TABLE 4 Structural parameter estimates: the best‐fitting nonlinear associative growth curve model
variable (DV) Independent variable
coefficient (b) Standard error
BAI intercept 24.73 1.04 <0.001
rTMS: Left versus right 2.30 1.31 0.07 0.08
Bilateral versus right −2.41 1.11 0.10 0.05
Left versus bilateral 4.71 1.37 0.16 0.003
Gender (female) 4.99 1.07 0.19 <0.001
BAI slope −8.13 0.79 <0.001
rTMS: Left versus right −0.38 1.00 −0.02 0.71
Bilateral versus right 1.13 0.79 0.07 0.23
Left versus bilateral −1.50 1.02 −0.08 0.22
Gender (female) −0.63 0.77 −0.04 0.42
HAMD intercept 22.14 0.46 <0.001
rTMS: Left versus right 3.19 0.57 0.32 <0.001
Bilateral versus right −2.35 0.50 −0.27 <0.001
Left versus bilateral 5.54 0.61 0.55 <0.001
Gender (female) 1.31 0.48 0.14 0.006
HAMD slope −8.61 0.56 <0.001
rTMS: Left versus right −2.20 0.69 −0.25 0.009
Bilateral versus right 0.85 0.59 0.11 0.48
Left versus bilateral −3.05 0.73 −0.35 <0.001
Gender (female) 0.26 0.55 0.03 0.64
Covariance Intercept and slope (BAI) −40.31 12.62 0.001
−0.32 5.37 0.95
Cross intercepts 19.53 2.95 <0.001
Cross slopes 14.97 2.46 <0.001
0.25 3.22 0.93
−2.95 2.08 0.15
Note. BAI: 21‐item Beck's Anxiety Inventory; HAMD: 17‐item Hamilton Depression Rating Scale; rTMS: repetitive transcranial magnetic stimulation; DV,
*pvalues adjusted for multiplicity (Multtest procedure; SAS institute, Cary, NC).
CHEN ET AL.
as the serotonin reuptake inhibitors and serotonin/adrenaline
reuptake inhibitors) infer therapeutic efficacy for both anxiety and
depression disorders. Theories exist that refer to multiple brain
regions contributing to the manifestation of anxiety, relative to
depressive disorders, which make focal targeting of brain regions
with rTMS practically difficult. rTMS applied to cortical regions
thought to play a role in modulating anxious experiences, such as the
dorsal anterior cingulate (dACC) and medial prefrontal regions (Etkin
& Schatzberg, 2011), have also not been systemically investigated.
With respect to the therapeutic potential of rTMS applied to
subcortical brain regions, electromagnetic pulses delivered by
conventional rTMS machines with figure‐of‐eight coils are not
engineered to penetrate to such depth below the cortical surface
(Paes et al., 2011), although this may explain the positive results deep
rTMS therapy have had in treating anxious depression (Kedzior et al.,
2015). Separate from the issue of neuroanatomical targets, the thus
far discouraging therapeutic outcomes of rTMS in anxiety disorders
may also relate to uncertainties in relation to optimal stimulation
parameters (Paes et al., 2011). rTMS therapeutic trials for anxiety
disorders have typically featured stimulation parameters analogous
with parameters used in depression trials when there is no evidence
to suggest mechanisms of therapeutic action, even if putative, are
transferrable from one group of disorders to another.
By way of limitations, heterogeneity in the rTMS protocols across
the three trials is noteworthy, particularly with reference to the
number of stimulation pulses applied per session and cumulatively
over a course of 20 sessions. Across the protocols, pulses applied per
session ranged between 900 and 5,625. The increased antidepressant
effect of LHF rTMS we observed could be attributable to the greater
number of pulses applied to patients randomised to this protocol.
There is no convincing evidence from published meta‐analyses,
however, that supports a clear correlation between the total number
of pulses applied through a course of rTMS and antidepressant
efficacy (Berlim, van den Eynde, et al., 2014; Herrmann & Ebmeier,
2006; Slotema et al., 2010). In other words, a therapeutic ceiling effect
could exist in rTMS, where more pulses do not equate to dramatically
superior antidepressant response. Rather, the superior antidepressant
effect we observed with LHF rTMS may be more likely attributable to
the higher stimulation intensities applied in these treatment arms
(120% RMT compared with 110% RMT). The incorporation of a
priming rTMS protocol (20 trains of sub‐RMT stimulation of 5‐second
duration at 6 Hz followed by 15 min of 1 Hz supra‐RMT stimulation,
both applied to the right DLPFC) also deserves consideration.
Published studies indicate this form of right‐sided rTMS may produce
superior antidepressant effects compared with RLF rTMS without
preceding priming stimulation (Fitzgerald et al., 2008; Iyer, Schleper, &
Wassermann, 2003) and thereby increased the overall therapeutic
efficacy in our priming RLF rTMS arm or across the sample. This is
unlikely, however, given patients randomised to the priming RLF
rTMS protocol made up 29.2% (91 of 312) of the total number of RLF
rTMS patients and 13.1% (91 of 697) of the total sample size. Given
the infinite combination of possibilities, the issue of heterogeneity
with stimulation parameters is frequently encountered in TMS
research (Lefaucheur et al., 2014). Despite minor heterogeneity in
the rTMS protocols we analysed, for the purpose of our study
objective, which was to investigate relative anxiolytic and antide-
pressant efficacy, our results are nonetheless likely to be clinically
informative. The fact that the studies were not sham‐controlled and
took place in inpatient settings are also worth noting. It is possible
that a placebo effect across the studies obscured between‐group
differences. Therapeutic measures available in an inpatient mental
health facility, such as regular clinician contact and availability of
therapeutic programs may have increased the magnitude of treatment
response in the populations studied. Again, for the purpose of
establishing relative anxiolytic and antidepressant superiority across
the three rTMS protocols, the absence of sham control and
investigation of an inpatient population are not likely to bias our
results. The choice to analyse treatment response at the week 4 time
point does, however, mean delayed treatment responses beyond this
time were not recorded. The decision to use week 4 as the treatment
endpoint was because this was the review time point standardised
across the three studies.
FIGURE 2 Mean reduction trends in BAI across the three rTMS
protocols. BAI: Beck's Anxiety Inventory; rTMS: repetitive
transcranial magnetic stimulation
FIGURE 3 Mean reduction trends in HAMD across the three
rTMS Protocols. HAMD: 17‐item Hamilton Depression Rating Scale;
rTMS: repetitive transcranial magnetic stimulation
CHEN ET AL.
In spite of the high comorbidity of anxiety and depression and the
prevalence of both disorders, research addressing whether certain
rTMS protocol(s) are particularly effective in anxiety symptoms
comorbid in depression is lacking. This study addresses this clinical
knowledge gap and is the largest analysis to date comparing the relative
therapeutic efficacy of three commonly used rTMS protocols. Our
findings suggest therapeutic equivalence across LHF rTMS, RLF rTMS,
and BL rTMS. It remains unclear why comorbid anxiety in depression
improves with rTMS, while its therapeutic efficacy in primary anxiety
disorders remains equivocal. Further research into the neurobiological
aetiology of both disorders and rTMS's mechanisms of action may
provide clues on how this technique may be harnessed for greater
therapeutic potential. Future randomised trials, prospectively evaluating
different rTMS protocols' effectiveness in treating anxiety symptoms in
depression can be of clinical value.
Leo Chen http://orcid.org/0000-0002-2371-8707
Baxter, A. J., Vos, T., Scott, K. M., Ferrari, A. J., & Whiteford, H. A. (2014).
The global burden of anxiety disorders in 2010. Psychological Medicine,
44(11), 2363–2374. https://doi.org/10.1017/S0033291713003243
Baxter, L. R., Jr., Schwartz, J. M., Phelps, M. E., Mazziotta, J. C., Guze, B. H.,
Selin, C. E., & Sumida, R. M. (1989). Reduction of prefrontal cortex
glucose metabolism common to three types of depression. Archives of
General Psychiatry,46(3), 243–250.
Bech, P., Kastrup, M., & Rafaelsen, O. J. (1986). Mini‐compendium of
rating scales for states of anxiety depression mania schizophrenia
with corresponding DSM‐III syndromes. Acta Psychiatrica Scandinavica.
Berlim, M. T., van den Eynde, F., Tovar‐Perdomo, S., & Daskalakis, Z.
(2014). Response, remission and drop‐out rates following high‐
frequency repetitive transcranial magnetic stimulation (rTMS) for
treating major depression: A systematic review and meta‐analysis of
randomized, double‐blind and sham‐controlled trials. Psychological
Berlim, M. T., McGirr, A., Beaulieu, M. M., & Turecki, G. (2011). High
frequency repetitive transcranial magnetic stimulation as an aug-
menting strategy in severe treatment‐resistant major depression: A
prospective 4‐week naturalistic trial. Journal of Affective Disorders,
130(1‐2), 312–317. https://doi.org/10.1016/j.jad.2010.10.011
Chen, J., Zhou, C., Wu, B., Wang, Y., Li, Q., Wei, Y., & Zou, D. (2013). Left
versus right repetitive transcranial magnetic stimulation in treating
major depression: A meta‐analysis of randomised controlled trials.
Psychiatry Research,210(3), 1260–1264.
Chen, J.‐J., Liu, Z., Zhu, D., Li, Q., Zhang, H., Huang, H., & Xie, P. (2014).
Bilateral vs. unilateral repetitive transcranial magnetic stimulation in
treating major depression: A meta‐analysis of randomized controlled
trials. Psychiatry Research,219(1), 51–57. https://doi.org/10.1016/j.
Crown, W. H., Finkelstein, S., Berndt, E. R., Ling, D., Poret, A. W., Rush, A.
J., & Russell, J. M. (2002). The impact of treatment‐resistant
depression on health care utilization and costs. Journal of Clinical
Curran, P. J., & Muthen, B. O. (1999). The application of latent curve
analysis to testing developmental theories in intervention research.
American Journal of Community Psychology,27(4), 567–595. https://doi.
Diefenbach, G. J., Bragdon, L., & Goethe, J. W. (2013). Treating anxious
depression using repetitive transcranial magnetic stimulation. Journal
of Affective Disorders,151(1), 365–368. https://doi.org/10.1016/j.jad.
Durmaz, O., Ebrinc, S., Ates, M. A., & Algul, A. (2017). Evaluation of
repetitive transcranial magnetic stimulation for treatment‐resistant
major depression and the impact of anxiety symptoms on outcome.
Psychiatry and Clinical Psychopharmacology,27(1), 14–18.
Etkin, A. (2012). Neurobiology of anxiety: From neural circuits to novel
solutions? Depression and Anxiety,29(5), 355–358.
Etkin, A., & Schatzberg, A. F. (2011). Common abnormalities and disorder‐
specific compensation during implicit regulation of emotional proces-
sing in generalized anxiety and major depressive disorders. American
Journal of Psychiatry,168(9), 968–978. https://doi.org/10.1176/appi.
Fitzgerald, P. B., Brown, T., Marston, N., Daskalakis, Z., de Castella, A., &
Kulkarni, J. (2003). Transcranial magnetic stimulation in the treatment
of depression: A double‐blind, placebo‐controlled trial. Archives of
General Psychiatry,60, 1002–1008.
Fitzgerald, P. B., Hoy, K., Gunewardene, R., Slack, C., Ibrahim, S., Bailey, M.,
& Daskalakis, Z. J. (2011). A randomized trial of unilateral and
bilateral prefrontal cortex transcranial magnetic stimulation in
treatment‐resistant major depression. Psychological Medicine,41(6),
Fitzgerald, P. B., Hoy, K., McQueen, S., Herring, S., Segrave, R., Been, G., &
Daskalakis, Z. J. (2008). Priming stimulation enhances the effective-
ness of low‐frequency right prefrontal cortex transcranial magnetic
stimulation in major depression. Journal of Clinical Psychopharmacology,
Fitzgerald, P. B., Hoy, K. E., Singh, A., Gunewardene, R., Slack, C., Ibrahim,
S., & Daskalakis, Z. J. (2013). Equivalent beneficial effects of unilateral
and bilateral prefrontal cortex transcranial magnetic stimulation in a
large randomized trial in treatment‐resistant major depression.
International Journal of Neuropsychopharmacology,16(9), 1975–1984.
George, M. S., Ketter, T., & Post, R. (1994). Prefrontal cortex dysfunction
in clinical depression. Depression,2,59–72.
George, M. S., Ketter, T. A., & Post, R. M. (1993). SPECT and PET imaging
in mood disorders. Journal of Clinical Psychiatry,54(Suppl), 6–13.
Pavlicova, M., & Sackeim, H. A. (2010). Daily left prefrontal transcranial
magnetic stimulation therapy for major depressive disorder: A sham‐
controlledrandomizedtrial.Archives of General Psychiatry,67(5), 507–
George, M. S., Nahas, Z., Kozol, F. A., Li, X., Yamanaka, K., Mishory, A., &
Bohning, D. E. (2003). Mechanisms and the current state of
transcranial magnetic stimulation. CNS Spectrums,8(7), 496–514.
Gorman, J. M. (1996). Comorbid depression and anxiety spectrum
disorders. Depression and Anxiety,4(4), 160–168. https://doi.org/10.
Greenberg, B. D., McCann, U. D., Benjamin, J., & Murphy, D. L. (1997).
Repetitive TMS as a probe in anxiety disorders: Theoretical
considerations and case reports. CNS Spectrums,2(1), 47–52.
Herrmann, L. L., & Ebmeier, K. P. (2006). Factors modifying the efficacy of
transcranial magnetic stimulation in the treatment of depression: A
review. Journal of Clinical Psychiatry,67(12), 1870–1876.
Hoy, K. E., Segrave, R. A., Daskalakis, Z. J., & Fitzgerald, P. B. (2012).
Investigating the relationship between cognitive change and anti-
depressant response following rTMS: A large scale retrospective
study. Brain Stimulation,5(4), 539–546. https://doi.org/10.1016/j.brs.
CHEN ET AL.
Iyer, M. B., Schleper, N., & Wassermann, E. M. (2003). Priming stimulation
enhances the depressant effect of low‐frequency repetitive transcra-
nial magnetic stimulation. Journal of Neuroscience,23(34), 10867–
Kedzior, K. K., Gellersen, H. M., Roth, Y., & Zangen, A. (2015). Acute
reduction in anxiety after deep transcranial magnetic stimulation
(DTMS) in unipolar major depression‐a systematic review and meta‐
analysis. Psychiatry Research,230(3), 971–974. https://doi.org/10.
Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., &
Walters, E. E. (2005). Lifetime prevalence and age‐of‐onset distribu-
tions of DSM‐IV disorders in the National Comorbidity Survey
Replication. Archives of General Psychiatry,62(6), 593–602. https://
Lefaucheur, J.‐P., André‐Obadia, N., Antal, A., Ayache, S. S., Baeken, C.,
Benninger, D. H., & Garcia‐Larrea, L. (2014). Evidence‐based guide-
lines on the therapeutic use of repetitive transcranial magnetic
stimulation (rTMS). Clinical Neurophysiology,125(11), 2150–2206.
Liston, C., Chen, A. C., Zebley, B. D., Drysdale, A. T., Gordon, R., Leuchter,
B., & Dubin, M. J. (2014). Default mode network mechanisms of
transcranial magnetic stimulation in depression. Biological Psychiatry,
76(7), 517–526. https://doi.org/10.1016/j.biopsych.2014.01.023
Machado, S., Paes, F., Velasques, B., Teixeira, S., Piedade, R., Ribeiro, P., &
Arias‐Carrion, O. (2012). Is rTMS an effective therapeutic strategy
that can be used to treat anxiety disorders? Neuropharmacology,62(1),
O'Reardon, J., Solvason, H., & Janicak, P. (2007). Efficacy and safety of
transcranial magnetic stimulation in the acute treatment of major
depression: A muiltisite randomized contorlled trial. Biological
Paes, F., Machado, S., Arias‐Carrión, O., Velasques, B., Teixeira, S., Budde,
H., & P Huston, J. (2011). The value of repetitive transcranial magnetic
stimulation (rTMS) for the treatment of anxiety disorders: An
integrative review. CNS & Neurological Disorders Drug Targets,10(5),
Pallanti, S., & Bernardi, S. (2009). Neurobiology of repeated transcranial
magnetic stimulation in the treatment of anxiety: A critical review.
International Clinical Psychopharmacology,24(4), 163–173. https://doi.
Pigot, M., Loo, C., & Sachev, P. (2008). Repetitive transcranial magnetic
stimulation as treatment for anxiety disorders. Expert Review of
Pridmore, S., Fernandes Filho, J. A., Nahas, Z., Liberatos, C., & George, M.
S. (1998). Motor threshold in transcranial magnetic stimulation: A
comparison of a neurophysiological method and a visualization of
movement method. The Journal of ECT,14(1), 25–27.
Ressler, K. J., & Mayberg, H. S. (2007). Targeting abnormal neural circuits
in mood and anxiety disorders: From the laboratory to the clinic.
Nature Neuroscience,10(9), 1116–1124. https://doi.org/10.1038/
Schutter, D. (2009). Antidepressant efficacy of high‐frequency transcra-
nial magnetic stimulation over the left dorsolateral prefrontal cortex
in double‐blind sham‐controlled designs: A meta‐analysis. Psychologi-
cal Medicine,39(31), 65–75.
Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J.,
Weiller, E., & Dunbar, G. C. (1998). The Mini‐International Neurop-
sychiatric Interview (M.I.N.I.): The development and validation of a
structured diagnostic psychiatric interview for DSM‐IV and ICD‐10.
Journal of Clinical Psychiatry,59(Suppl 20), 22–33.
Slotema, C., Blom, J., Hoek, H., & Sommer, I. (2010). Should we expand the
toolbox of psychiatric treatment methods to include Repetitive
Transcranial Magnetic Stimulation (rTMS)? A meta‐analysis of the
efficacy of rTMS in psychiatric disorders. Journal of Clinical Psychiatry,
Thase, M. E., & Rush, A. J. (1997). When at first you don't succeed:
Sequential strategies for antidepressant nonresponders. Journal of
Clinical Psychiatry,58(Suppl 13), 23–29.
Tik, M., Hoffmann, A., Sladky, R., Tomova, L., Hummer, A., Navarro de Lara,
L., & Windischberger, C. (2017). Towards understanding rTMS
mechanism of action: Stimulation of the DLPFC causes network‐
specific increase in functional connectivity. NeuroImage,162, 289–
Whiteford, H. A., Degenhardt, L., Rehm, J., Baxter, A. J., Ferrari, A. J.,
Erskine, H. E., & Johns, N. (2013). Global burden of disease
attributable to mental and substance use disorders: Findings from
the Global Burden of Disease Study 2010. The Lancet,382(9904),
Zwanzger,P.,Fallgatter,A.J.,Zavorotnyy, M., & Padberg, F. (2009).
Anxiolytic effects of transcranial magnetic stimulation—an alter-
native treatment option in anxiety disorders? Journal of Neural
Transmission,116(6), 767–775. https://doi.org/10.1007/s00702‐
How to cite this article: Chen L, Hudaib AR, Hoy KE,
Fitzgerald PB. Is rTMS effective for anxiety symptoms in
major depressive disorder? An efficacy analysis comparing
left‐sided high‐frequency, right‐sided low‐frequency, and
sequential bilateral rTMS protocols. Depress Anxiety. 2019;
CHEN ET AL.