Brain Stimulation 8 (2015) 801e807

Contents lists available at ScienceDirect

Brain Stimulation
journal homepage: [Link]

Treatment of Chronic Facial Pain Including Cluster Headache by

Repetitive Transcranial Magnetic Stimulation of the Motor Cortex
With Maintenance Sessions: A Naturalistic Study
Hasan Hodaj a, *, Jean-Pierre Alibeu a, Jean-François Payen a, Jean-Pascal Lefaucheur b
a
Centre de la douleur, pôle anesthésie-réanimation, CHU de Grenoble, Grenoble, France
b
Service de Physiologie, Explorations Fonctionnelles, Hôpital Henri Mondor, Université Paris Est Créteil, EA 4391, Créteil, France

a r t i c l e i n f o a b s t r a c t

Article history: Objective: To assess the long-term maintenance of analgesia induced by high-frequency repetitive
Received 13 June 2014 transcranial magnetic stimulation (rTMS) of the motor cortex contralateral to pain in a naturalistic study
Received in revised form of patients with chronic refractory facial pain.
24 January 2015
Methods: 55 patients were included (cluster headache, n ¼ 19; trigeminal neuropathic pain, n ¼ 21;
Accepted 31 January 2015
atypical facial pain, n ¼ 15). The rTMS protocol consisted of an “induction phase” of one daily rTMS
Available online 13 May 2015
session for five days per week during two consecutive weeks, followed by a “maintenance phase” of two
sessions during one week, then one session in weeks 4 and 6, and a monthly session for the next five
Keywords:
Chronic pain
months. In a subset of patients, navigated targeting was performed and session duration was shortened
Cluster headache from 20-min to 10-min (with the same number of 2000 pulses per session). The analgesic effect of rTMS
Facial pain was assessed on a 0e10 visual numerical scale from 15 to 180 days after treatment initiation.
rTMS Results: All pain measures significantly decreased from baseline to D15: the intensity of permanent pain
Trigeminal neuralgia (5.2  1.6 to 3.2  1.9) and paroxysmal pain (8.6  1.5 to 4.5  3.4), as well as the daily number of painful
attacks (5.6  3.1 to 2.3  3.1). The percentage of responders (defined as pain score decrease 30%) was
73% at D15 and dropped to 40% at D180. The analgesic effect was similar regardless of the type of pain
and was significantly lower when session duration was shortened, irrespective of the number of pulses.
Conclusion: This long-term maintenance rTMS protocol can be a therapeutic option in the clinical
management of patients with chronic refractory facial pain, including cluster headache. However, only
part of the patients respond to this technique and session duration should not be reduced.
Ó 2015 Elsevier Inc. All rights reserved.

Introduction data concern migraine, targeting the occipital visual cortex [16],
including with single-pulse TMS approach [17], the dorsolateral
Following the first reports of analgesia produced by surgically- prefrontal cortex [18], or the primary motor cortex, which seems to
implanted epidural motor cortex stimulation in the early nineties be a good target [19,20]. Regarding facial neuropathic pain, beyond
[1e4], facial pain was usually considered a good indication of this single cases reported separately [21] or included in large series of
technique [5,6]. More recently, repetitive transcranial magnetic patients with pain of various origins [22,23], the main data consist
stimulation (rTMS) of the motor cortex, a non-invasive technique of of six controlled studies of 7e24 patients [24e29]. In these studies,
cortical stimulation, was also used to relieve refractory neuropathic facial pain was mostly secondary to surgical or traumatic lesion of
pain [7e15]. However, only few studies assessed the analgesic effect the trigeminal nerve and high-frequency rTMS was delivered over
of rTMS specifically on facial pain or headache disorders. Most rTMS the motor cortical area corresponding to the painful face [24,26,29]
or to the hand of the painful side [25e28]. These trials were based
on single rTMS sessions, except two studies in which patients were
Declaration of interest: The authors report no conflict of interest. stimulated for five consecutive days [28,29]. These two series
The rTMS equipment was supported in part by the “Fondation APICIL.” showed significant analgesic effects lasting for 2 weeks after the
* Corresponding author. Centre de la douleur, pôle anesthésie-réanimation,
end of rTMS sessions. One series included only patients with facial
CHU de Grenoble, BP217, 38043 Grenoble, France. Tel.: þ33 476765213;
fax: þ33 476765951. pain [29] and the other included patients with trigeminal neuralgia
E-mail address: HHodaj@[Link] (H. Hodaj). or post-stroke pain at the face and upper limb, with no obvious

[Link]
1935-861X/Ó 2015 Elsevier Inc. All rights reserved.
802 H. Hodaj et al. / Brain Stimulation 8 (2015) 801e807

differences in rTMS efficacy between these two clinical conditions pain (AFP) according to the ICHD, 1st edition (persistent facial pain
[28]. Some studies showed a higher percentage of responders in that does not have the characteristics of other cranial neuralgias
cases of facial pain compared to other pain sites [25,26], while this and is not associated with physical signs or a demonstrable organic
difference failed to reach significance in other studies [24,27]. In the cause) [39]. The following versions of the ICHD have adopted the
present study, we have evaluated the analgesic efficacy of rTMS not term “persistent idiopathic facial pain” (PIFP) [38], but the term AFP
only in painful trigeminal neuropathy or neuralgia, but also in was preferred as it continues to be commonly used in other clas-
other types of facial pain syndromes refractory to conventional sifications [40] or by clinicians [41].
treatment. All patients with CH had previously received the reference
Cortical stimulation produces analgesia through various treatments in this setting, including verapamil (n ¼ 15), lithium
possible mechanisms, including the activation of neural structures (n ¼ 6), and occipital nerve infiltration (n ¼ 8) for attack prevention
at a distance from the site of stimulation [30]. For example, motor on the one hand, and oxygen therapy (n ¼ 17) and sumatriptan
cortex stimulation can reduce hyperactivity of thalamic relays (n ¼ 12) for acute attacks on the other hand. They had also received
involved in the transmission of painful stimuli [1,31]. Regarding the analgesic treatments that are less specifically indicated in this dis-
involved neurotransmitters, analgesia could result from the mod- order (including antiepileptics, n ¼ 9 and antidepressants, n ¼ 7).
ulation of endogenous opioidergic control [32,33] or the restoration Antiepileptic drugs (mainly carbamazepine) were also adminis-
of intracortical gabaergic inhibition [34]. In the French [35] and tered in all patients with TNP or AFP, associated with antidepres-
international [36] recommendations for the use of rTMS, experts sants in 29 patients (81%) and trigeminal ganglion block in 15
retained a level A of evidence regarding the analgesic effect of high- patients (42%).
frequency motor cortex rTMS in patients with chronic neuropathic
pain. However, the statistical difference between the analgesic ef- rTMS protocol
fects produced by active and sham rTMS does not necessarily meet
the threshold of clinical significance [37]. Further studies are still Stimulation was performed using a MagPro stimulator (Mag-
needed before considering rTMS in the therapeutic armamen- Venture, Farum, Denmark) with a dynamic cooled figure-of-eight
tarium against pain, especially to assess the value of this technique coil. The stimulation target was the motor cortical representation
to produce a real clinical benefit in the long term. of the face, contralateral to the painful side, according to motor
Thus, our goal was not to demonstrate the actual analgesic ef- evoked potential (MEP) recordings. A non-navigated procedure was
ficacy of motor cortex rTMS compared to placebo, as it was already performed in 33 patients, whereas from January 2011, in 22 pa-
done, but to evaluate the benefits of this treatment in “real life”, tients, the site of cortical stimulation was determined using a TMS
namely clinical practice, in which placebo stimulations are not Navigator system (Localite, Sankt Augustin, Germany).
performed. This is the reason why we designed a naturalistic study Stimulation was performed at 10 Hz with an intensity set at 80%
of rTMS therapy for pain, which also addressed the issue of pro- of the resting motor threshold (RMT) determined as usual [42],
longing rTMS-induced analgesia in the long term by means of using the MEP monitor amplifier of the MagPro stimulator. Between
maintenance sessions. In addition, the influence of rTMS session January 2008 and October 2010, each rTMS session consisted of 40
duration and image-guided navigation on rTMS efficacy has been trains of 5-sec duration with intertrain interval (ITI) of 25 s, leading
studied. to deliver 2000 pulses in 20 min (20-min session) (22 patients).
From November 2010, session duration was shortened from 20 to
Methods 10 min (10-min session) (33 patients), consisting of 20 trains of 10-
sec duration with ITI of 20 s (15 patients, until March 2012), and
Patients then 40 trains of 5-sec duration with ITI of 10 s (18 patients, from
April 2012). The therapy protocol consisted of an “induction phase”
In this open-label, naturalistic study, 55 patients (30 men and 25 of one session per day for five days during two consecutive weeks
women) underwent motor cortex rTMS for the treatment of chronic (weeks 1 and 2), then 2 sessions in the next week (week 3) for a
facial pain between January 2008 and January 2014. Pain was pre- total of 12 sessions. Then, in patients with clinical response, defined
sent for more than one year (or more than 6 months in case of as a decrease in pain score 30% on a 0e10 visual numerical scale
postherpetic neuralgia) and was refractory to conventional therapy (VNS), a maintenance therapy was undertaken, consisting of one
or associated with poor drug tolerance. In 19 patients, facial pain session during weeks 4 and 6, and then a monthly session for the
met the criteria for diagnosis of cluster headache (CH) according to next five months, for a total of 7 sessions.
the International Classification of Headache Disorders (ICHD), 3rd
edition, of the International Headache Society [38]. In 10 patients, Clinical assessment
facial pain was secondary to a surgical or traumatic injury of the
trigeminal nerve or ganglion (traumatic trigeminal neuropathic Patients were assessed at baseline (before rTMS therapy) and 15,
pain, traTNP), including surgical treatment of classical trigeminal 30, 90, and 180 days after treatment initiation (D15, D30, D90,
neuralgia by percutaneous radiofrequency thermocoagulation, D180). In case of permanent pain, the average daily pain intensity
n ¼ 4, microvascular decompression, n ¼ 2, or percutaneous balloon was scored on a 0e10 VNS. In case of paroxysmal pain, the average
compression, n ¼ 1; posterior fossa (neuroma) surgery, n ¼ 2; pain intensity during attacks was also scored on a 0e10 VNS and the
traumatic trigeminal nerve injury, n ¼ 1. In 11 patients, facial pain number of painful attacks per day was recorded. The average per-
was secondary to an inflammatory or infectious lesion of the tri- centage of change for both permanent and paroxysmal pain was
geminal nerve or nuclei (inflammatory trigeminal neuropathic calculated following rTMS therapy compared to baseline at each
pain, infTNP), including herpes zoster infection, n ¼ 4; other viral time point, the result obtained at D15 being the primary endpoint of
infection, n ¼ 1; chronic inflammatory disease, such as Sjögren’s the study. Patients were classified into four groups according to
syndrome, n ¼ 3; brainstem lesion in the context of multiple scle- their analgesic response [3]: very good response (pain reduction
rosis, n ¼ 3. In 15 patients, facial pain was related to unclear 70%), good response (pain reduction from 50% to 69%), moderate
pathophysiology in the context of dental surgery, n ¼ 7; radio- response (pain reduction from 30% to 49%), and poor or no response
therapy for meningioma, n ¼ 1; stroke, n ¼ 1; undetermined cause, (pain reduction <30%). Finally, the global clinical effect of rTMS
n ¼ 6. These cases met the criteria for diagnosis of atypical facial therapy was self-assessed by each patient at D90 on the Clinical
 H. Hodaj et al. / Brain Stimulation 8 (2015) 801e807 803

Global Impression of Change (CGI-C) scale [43], ranging from 1: Paroxysmal pain was present in all patients with CH and 30 patients
‘very much improved’ to 7: ‘very much worse.’ At each time point of (83%) with TNP/AFP, but average pain intensity also did not differ
assessment (D15, D30, D90, D180), patients with poor or no between the two groups (8.5  1.8 versus 8.6  1.3, P ¼ 0.36).
response were withdrawn from the study and rTMS therapy was Finally, patients with CH had less painful attacks per day than pa-
stopped for them. tients with TNP/AFP (3.5  2.2 versus 6.9  2.9, P ¼ 0.0001).

Statistical analysis Analgesic effect of rTMS

First, differences in terms of gender, age, disease duration, RMT, The rTMS therapy was well tolerated without any serious
and baseline pain measures (permanent and paroxysmal pain in- adverse events. All the 55 initially included patients have
tensity and number of attacks) according to pain type (CH versus completed the “induction phase” of the protocol. As previously
TNP/AFP) were studied using the Fisher and ManneWhitney tests. mentioned, patients with poor or no response were excluded from
Second, the analgesic effect of rTMS was assessed by comparing the study at each time point of assessment (D15, n ¼ 15; D30, n ¼ 5,
pain measures (permanent and paroxysmal pain intensity and D90, n ¼ 3; D180, n ¼ 4). In addition, 6 patients were finally lost to
number of attacks) between baseline and D15 using the Wilcoxon follow-up. Thus, 22 patients who were responders to rTMS have
test. This analysis was performed in the entire series of patients and completed the “maintenance phase” of the protocol (CH, n ¼ 5/19,
in the various groups of patients according to pain type (CH, traTNP, aTNP, n ¼ 2/10, infTNP, n ¼ 7/11, and AFP, n ¼ 8/15).
infTNP, and AFP). For the patients who completed the study to From baseline to D15, the intensity of ongoing pain decreased in
D180, the influence of time on pain measures was studied using the patients with permanent pain (5.2  1.6 to 3.2  1.9, P < 0.0001,
Friedman and Wilcoxon tests (comparing the entire data set from Wilcoxon test). Pain intensity during acute attacks also decreased
D15 to D180 and only D15 versus D180 data, respectively). The re- (8.6  1.5 to 4.5  3.4, P < 0.0001), as well as the number of painful
sults were also analyzed in terms of responders (defined as having attacks per day (5.6  3.1 to 2.3  3.1, P < 0.0001). The analgesic
pain reduction 30% compared to baseline) using the Chi-squared effect of rTMS at D15 was significant on all pain measures in all
test. groups of patients according to pain type (CH, traTNP, infTNP, and
Third, analyses were performed in the entire series of patients to AFP), except for permanent pain in patients with CH (P ¼ 0.0625)
determine whether some factors were associated with the anal- (Fig. 1).
gesic effect of rTMS, defined as the percentage of change in pain In the 22 patients who completed the study to D180, we found
measures (permanent and paroxysmal pain intensity and number significant changes between D15 and D180 regarding all pain
of attacks) between pre-rTMS baseline and D15. The influence of measures (Fig. 2). Pairwise comparisons between D15 and D180
navigation and session duration was studied using the Kruskale showed a further reduction of the intensity of permanent pain at
Wallis and ManneWhitney tests by comparing three groups of the end of follow-up (2.8  1.5 to 2.0  1.5, P ¼ 0.0085, Wilcoxon
patients: (i) non-navigated 20-min rTMS session (22 patients); (ii) test). Results were not analyzed according to pain type because of
non-navigated 10-min rTMS session (11 patients); (iii) navigated the small number of patients in each subgroup.
10-min rTMS session (22 patients). The influence of age, disease Although the analgesic effect tended to improve over time in the
duration, RMT, and baseline pain measures (permanent and patients who completed the maintenance phase of rTMS therapy,
paroxysmal pain intensity and number of attacks) was studied us- the percentage of responders (defined as having pain reduction
ing the Spearman test. The influence of gender was studied using 30% compared to baseline) decreased from 73% (40 patients) at
the ManneWhitney test. The influence of pain type (CH versus TNP/ D15 to 40% (22 patients) at D180 in our series of patients (P ¼ 0.005,
AFP and CH versus traTNP, infTNP, and AFP) was studied using the Chi-squared test). The percentage of responders remained stable in
ManneWhitney and KruskaleWallis tests. The influence of medi- patients with infTNP (from 64% to 64% (7 patients), P ¼ 1.000) and
cation (antiepileptic drugs or not and antidepressants or not) was decreased, but not to a significant level, in patients with AFP (from
studied using the ManneWhitney test. 87% (13 patients) to 53% (8 patients), P ¼ 0.189), traTNP (from 70% (7
In all cases, the significance level of P value was set at 0.05. patients) to 20% (2 patients), P ¼ 0.116), and CH (from 68% (13 pa-
Statistical analyses were performed using the Stata 11.0 (StataCorp, tients) to 26% (5 patients), P ¼ 0.066). The rTMS response over time
College Station, Texas, USA) and InStat 3 (GraphPad Software, San is presented in details (very good, good, moderate, or poor or no
Diego, CA, USA) softwares. Data are presented as mean  standard response) in Table 1 for the entire series of patients and according to
deviation. each pain type.
Regarding the CGI-C score at D90, 10 patients felt ‘very much
Results improved’ (score 1, 18%) (CH, n ¼ 4/19, aTNP, n ¼ 3/10, infTNP, n ¼ 1/
11, and AFP, n ¼ 2/15), 13 patients felt ‘much improved’ (score 2,
Baseline differences according to pain type 24%), 9 patients felt ‘moderately improved’ (score 3, 16%), 22 pa-
tients felt ‘unchanged’ (score 4, 40%), and 1 patient felt ‘moderately
Patients with CH were younger than patients with TNP/AFP worsened’ (score 5, 2%). At the same time point (D90), from the
(45.6  13.4 versus 66.1  11.8 years; P < 0.0001, ManneWhitney entire series of patients, 10 patients (18%) stopped all drug treat-
test), but the duration of pain was similar (10.5  10.2 versus ment and 14 patients (25%) frankly reduced drug treatment. In
7.6  8.5 years, P ¼ 0.24), as well as the RMT (52.7  8.3 versus addition, from the 12 patients with CH initially treated by suma-
54.6  10.1% of maximal stimulator output, P ¼ 0.42). There was triptan, eight patients were sufficiently relieved to stop or reduced
also a gender difference between groups, a majority of patients with this treatment at D15. Five of these patients remained free of drug
CH being men whereas a majority of patients with TNP/AFP being at D30 and three at D90.
women (male/female sex-ratio: 5.33 versus 0.64; Fisher test,
P ¼ 0.002). Factors associated with rTMS analgesic effect
At baseline, before rTMS therapy, permanent pain was present in
8 patients (42%) with CH and 31 patients (86%) with TNP/AFP, but First, we found an overall influence of the change in rTMS pro-
average pain intensity did not differ between the two groups tocol regarding session duration and navigation use on the anal-
(4.8  1.6 versus 5.3  1.6, P ¼ 0.36, ManneWhitney test). gesic effect of rTMS between baseline and D15 for all pain measures
804 H. Hodaj et al. / Brain Stimulation 8 (2015) 801e807

Figure 1. Mean (þSD) intensity of permanent and paroxysmal pain on a 0e10 visual numerical scale and daily number of painful attacks at baseline and 15 days (D15) after rTMS
therapy initiation in the 55 patients who were initially enrolled. P values of pairwise comparison between baseline and D15 are presented (Wilcoxon test).

(Fig. 3). To distinguish between the respective effect of session paroxysmal pain intensity and number of attacks, respectively),
duration and navigation, we compared 20-min versus 10-min non- RMT (P ¼ 0.07, 0.07, and 0.32), baseline permanent pain intensity
navigated rTMS on the one hand, and non-navigated versus navi- (P ¼ 0.82, 0.40, and 0.56), paroxysmal pain intensity (P ¼ 0.30, 0.37,
gated 10-min rTMS on the other hand. Session duration shortening and 0.56), number of attacks (P ¼ 0.67, 0.69, and 0.43), gender
decreased the analgesic effect of non-navigated rTMS, at least (P ¼ 0.41, 0.29, and 0.38, ManneWhitney test), or pain type
regarding permanent pain (54% versus 25% mean pain relief for 20- (P ¼ 0.78, 0.33, and 0.58 for CH versus TNP/AFP and P ¼ 0.70, 0.79,
min and 10-min session duration, respectively, P ¼ 0.0395, and 0.64 for CH versus traTNP, infTNP, and AFP, KruskaleWallis
ManneWhitney test). In contrast, the use of navigation did not test). There was also no influence of medication on the analgesic
influence the analgesic effect of 10-min rTMS whatever the pain effect of rTMS, regarding either antiepileptic drugs (P ¼ 0.85, 0.71,
measure (P > 0.50). and 0.15, ManneWhitney test) or antidepressants (P ¼ 0.17, 0.29,
Second, we found an influence of age on the analgesic effect of and 0.69).
rTMS between baseline and D15. There was a positive correlation
between age and the effect of rTMS on permanent pain intensity Discussion
(r ¼ 0.33, P ¼ 0.04, Spearman test), paroxysmal pain intensity
(r ¼ 0.34, P ¼ 0.02), and number of attacks (r ¼ 0.31, P ¼ 0.03). This naturalistic study shows that rTMS can be used to signifi-
In contrast, we did not find an influence of disease duration cantly reduce facial pain of various origins for at least 6 months in
(P ¼ 0.45, 0.08, and 0.21 for the effect on permanent and responders. At the end of the induction phase, the intensity of

Figure 2. Mean (þSD) intensity of permanent and paroxysmal pain on a 0e10 visual numerical scale and daily number of painful attacks from baseline (D0) to 180 days (D180) after
rTMS therapy initiation in the 22 patients who completed the study. P values of ANOVA with repeated measures for the entire data set are presented on the left (Friedman test).
P values of pairwise comparison between D15 and D180 are presented in italics on the right (Wilcoxon test).
 H. Hodaj et al. / Brain Stimulation 8 (2015) 801e807 805

Table 1 pronounced in patients with CH. However, in responders who

Number of patients according to the level of pain reduction at 15e180 days after the completed the maintenance phase of rTMS therapy, the analgesic
initiation of repetitive transcranial magnetic stimulation (rTMS) therapy.
effect tended to improve over time, especially regarding permanent
Very good Good Moderate Poor or no Withdrawn Lost to pain relief. This means that there are good responders, for whom
response response response response from rTMS follow-up rTMS can provide a lasting improvement, relevant in terms of
Entire series of patients (n ¼ 55) clinical management. After six months of treatment, 17 patients
D15 16 16 8 15 0 0
(31% of the entire series) with drug-resistant chronic facial pain still
D30 12 13 10 5 15 0
D90 10 13 9 3 20 0 have good pain relief (50%), including 5 patients with CH (26% of
D180 9 8 5 4 23 6 this group). The number of patients amounts to 22 (40% of the
Patients with cluster headache (n ¼ 19) entire series) when considering an improvement of 30%.
D15 6 5 2 6 0 0 This result is encouraging, based on a single rTMS session per
D30 5 4 1 3 6 0
D90 4 3 1 2 9 0
month as maintenance therapy. In a study on fibromyalgia [44], it
D180 4 1 0 0 11 3 has been suggested that when moving from fortnightly to monthly
Patients with traumatic trigeminal neuropathic pain (n ¼ 10) stimulation the analgesic effect could decrease. So, the present
D15 2 4 1 3 0 0 finding of a reduction of the number of responders over time could
D30 2 3 1 1 3 0
well be related to a too long interval between consecutive rTMS
D90 3 2 1 0 4 0
D180 1 1 0 2 4 2 sessions and not to a loss of rTMS effect per se. Therefore, increasing
Patients with inflammatory trigeminal neuropathic pain (n ¼ 11) the number of sessions during the maintenance phase is desirable
D15 4 1 2 4 0 0 and could perhaps afford to maintain in more patients the analgesic
D30 1 4 2 0 4 0 effect obtained during the induction phase. This deserves
D90 1 4 2 0 4 0
D180 2 3 2 0 4 0
confirmation.
Patients with atypical facial pain, idiopathic or of central cause (n ¼ 15) The present study has several limitations. First, it is a retro-
D15 4 6 3 2 0 0 spective, non-controlled study that does not allow us to distinguish
D30 4 2 6 1 2 0 between real and placebo effects. In particular, we cannot rule out a
D90 2 4 5 1 3 0
placebo effect based on the expectation of a “promised” effect, as
D180 2 3 3 2 4 1
previously reported [45], even if the enrolled patients had pain that
Very good response: pain reduction 70%; good response: pain reduction from 50% was previously resistant to all types of treatment. Nevertheless, this
to 69%; moderate response: pain reduction from 30% to 49%; poor or no response:
pain reduction <30%.
naturalistic study has the merit of considering rTMS therapy in the
“real world” of clinical practice, which of course would not include
sham-control stimulations. The other limitations are the relative
permanent pain was decreased by about 2 points on a 0e10 VNS heterogeneity of patients’ profile (regarding pain origin and pre-
(39% reduction in average), the intensity of acute pain during at- vious treatments) and the simple evaluation method based on the
tacks was reduced by about 4 points (47%), and the number of VNS (without multi-dimensional assessment of pain or quantifi-
painful attacks per day was reduced by more than 3 (59%). The cation of drug consumption changes). This study also lacks of
analgesic effect was quite similar in all groups of patients, regard- control measurements, such as multiple baseline assessments or
less of the type of pain, with a percentage of responders at D15 comparison to historical pain ratings, which could help to more
(defined as a decrease in pain score 30%) ranging between 64% accurately allocate the effects to the intervention rather than con-
and 87% (73% in average). At D90, 58% of patients still reported to be founds. Finally, it should be noted that the methods were empiri-
improved on CGI-C scale and 43% reduced or stopped analgesic cally modified during the study period, but this allowed us to
drug treatment. The overall rate of responders decreased from 73% demonstrate a deleterious impact of shortening rTMS session
at D15 to 40% at D180, the decrease being perhaps a little more duration on the analgesic effect, while the use of navigation had no
impact.
In 60% of our patients, session duration was reduced from
20 min to 10 min, with a shortened value of ITI, remaining in the
safety limits of rTMS guidelines [34,46]. A shortened session
duration potentially offers various advantages in practice for both
operators and patients, including treatment tolerability, even for
higher doses per session [47,48]. However, we found that short-
ening session duration could have led to lose the therapeutic
benefit of rTMS, even if the number of pulses was kept constant
(2000 pulses). Actually, the percentage of responders at D180 was
55% in the group of patients who received 20-min rTMS sessions,
but only 30% in the group of patients who received 10-min rTMS
sessions. To our knowledge, this finding is original and has not been
previously reported. Other studies have shown that shorter rTMS
sessions had a lower clinical impact, but these sessions also
included fewer stimuli. For example, in 2007, one large multicenter
trial reported a lack of antidepressant efficacy of high-frequency
Figure 3. Mean percentage of reduction between baseline and 15 days (D15) after
rTMS therapy initiation regarding all pain measures (intensity of permanent and rTMS of the dorsolateral prefrontal cortex compared to placebo
paroxysmal pain on a 0e10 visual numerical scale and daily number of painful attacks), [49], whereas another large multicenter trial with a similar
according to session duration and navigation use. From left to right are presented: P approach was positive [50]. Among the several methodological
values of ANOVA with for the entire data set (KruskaleWallis test) and P values of differences that could explain the absence of efficacy in the first of
comparison between 20-min and 10-min non-navigated rTMS condition (underlined
these two studies, there was a shorter session duration (16.6 versus
italics) on one hand and between non-navigated and navigated 10-min rTMS (italics)
on the other hand (ManneWhitney test). The number of patients in the three groups 37.5 min), but also a smaller number of pulses per session (2000
was 22, 11, and 22, respectively. versus 3000). Our study suggests that the duration of daily rTMS
806 H. Hodaj et al. / Brain Stimulation 8 (2015) 801e807

sessions should not be reduced to 10 min, even with a high number reported in patients with fibromyalgia [44,48,59] and few case re-
of pulses per session. Interestingly, in experiments performed in the ports of neuropathic pain [21,60]. Our study supports the use of
rat, the induction of gene transcription in basal ganglia following rTMS in the “real world” of clinical practice for treating chronic
cortical stimulation was found to depend on the total number of facial pain. Furthermore, the analgesic effect might be enhanced by
shocks delivered but also on the total duration of stimulation prolonging the duration of each rTMS session (20 min for all pa-
[51,52]. tients) and by intensifying the protocol of maintenance therapy
Regarding navigation, we did not find any difference in rTMS (more than a monthly session).
efficacy whether the procedure was navigated or not, at least for 10- These encouraging results remain to be confirmed by a pro-
min session duration. However, we have not checked whether spective sham-controlled study to appraise the impact of the pla-
navigation resulted in a change in the location of the target. The cebo effect. Efforts should be made to improve the maintenance
non-navigated procedure targeted the facial motor hotspot, which protocol to keep constant the rate of responders. Two methodo-
has been shown to be not the optimal procedure for facial pain [26]. logical issues also need to be further investigated: the optimal
This issue deserves to be further studied, because ambiguity per- target location for facial pain and the respective influence of session
sists about the best target for motor cortex rTMS-induced analgesia duration and pulse number on the analgesic effect. The repetition of
according to pain location. Anyway, one advantage of the navigated daily sessions during an induction phase, followed by weekly or
procedure was to target the anatomical representation of the face fortnightly sessions during a maintenance phase opens promising
on the precentral gyrus with no longer need to record facial MEPs. perspectives for the technique of rTMS as an alternative treatment
In addition, navigation is known to improve the reliability of coil of chronic pain.
repositioning from one session to another [53].
Finally, the only other factor that we found influencing the re-
sults was the age of the patient: curiously, better results were found
References
in older patients. The fact that patients with CH were younger than
the others may have an influence, but this is unlikely because [1] Tsubokawa T, Katayama Y, Yamamoto T, Hirayama T, Koyama S. Treatment of
analgesic effects of rTMS did not significantly differ with pain type. thalamic pain by chronic motor cortex stimulation. Pacing Clin Electrophysiol
1991;14:131e4.
Anyway, this should be acknowledged and assessed in further
[2] Meyerson BA, Lindblom U, Linderoth B, Lind G, Herregodts P. Motor cortex
studies. In contrast, pain intensity at baseline, disease duration, stimulation as treatment of trigeminal neuropathic pain. Acta Neurochir Suppl
RMT (which determines the intensity of stimulation), or pain type (Wien) 1993;58:150e3.
did not influence the analgesic effect of rTMS. [3] Fontaine D, Hamani C, Lozano A. Efficacy and safety of motor cortex stimu-
lation for chronic neuropathic pain: critical review of the literature.
As mentioned in the introduction, rTMS has not been frequently J Neurosurg 2009;110:251e6.
studied in patients with facial pain until now. In addition, literature [4] Nizard J, Raoul S, Nguyen JP, Lefaucheur JP. Invasive stimulation therapies for
data, summarized in the introduction, mostly concerned patients the treatment of refractory pain. Discov Med 2012;14:237e46.
[5] Nguyen JP, Lefaucheur JP, Raoul S, Roualdes V, Péréon Y, Keravel Y. Traitement
with TNP and a few patients with AFP [26,29]. An originality of our des douleurs trigéminales neuropathiques par stimulation corticale. Neuro-
study was to apply motor cortex rTMS in refractory CH, which had chirurgie 2009;55:226e30.
not been previously reported to our knowledge. Refractory CH is a [6] Cruccu G, Aziz TZ, Garcia-Larrea L, et al. EFNS guidelines on neurostimulation
therapy for neuropathic pain. Eur J Neurol 2007;14:952e70.
very disabling condition that can lead to the indication of invasive [7] Lefaucheur JP, Drouot X, Keravel Y, Nguyen JP. Pain relief induced by repetitive
procedures, such as hypothalamic or occipital nerve stimulation transcranial magnetic stimulation of precentral cortex. Neuroreport
[54]. These procedures are based on pathophysiological hypotheses 2001;12:2963e5.
[8] Lefaucheur JP. The use of repetitive transcranial magnetic stimulation (rTMS)
involving connections between the hypothalamus or the great oc- in chronic neuropathic pain. Neurophysiol Clin 2006;36:117e24.
cipital nerve and the spinal trigeminal nucleus [55e57]. The anal- [9] Fregni F, Freedman S, Pascual-Leone A. Recent advances in the treatment of
gesic effect of rTMS could rely on the activation of descending chronic pain with non-invasive brain stimulation techniques. Lancet Neurol
2007;6:188e91.
inhibitory controls to the trigeminal nuclei in the brainstem.
[10] Lefaucheur JP. Use of repetitive transcranial magnetic stimulation in pain re-
Whatever the mechanisms involved, the use of rTMS seems lief. Expert Rev Neurother 2008;8:799e808.
promising, as a non-invasive treatment option before considering [11] Lefaucheur JP, Antal A, Ahdab R, et al. The use of repetitive transcranial
invasive techniques in CH. However, it is unlikely that the response magnetic stimulation (rTMS) and transcranial direct current stimulation
(tDCS) to relieve pain. Brain Stimul 2008;1:337e44.
to motor cortex rTMS can be used as a predictor of the outcome of [12] Leung A, Donohue M, Xu R, et al. rTMS for suppressing neuropathic pain: a
occipital nerve stimulation, as is the case for epidurally implanted meta-analysis. J Pain 2009;10:1205e16.
motor cortex stimulation [58]. [13] Nizard J, Lefaucheur JP, Helbert M, de Chauvigny E, Nguyen JP. Non-invasive
stimulation therapies for the treatment of refractory pain. Discov Med
2012;14:21e31.
Conclusion [14] Plow EB, Pascual-Leone A, Machado A. Brain stimulation in the treatment of
chronic neuropathic and non-cancerous pain. J Pain 2012;13:411e24.
[15] Lefaucheur JP. Pain Handb Clin Neurol 2013;116:423e40.
The present study shows that with a maintenance protocol, [16] Brigo F, Storti M, Nardone R, et al. Transcranial magnetic stimulation of visual
rTMS therapy can be efficacious to control pain for several months cortex in migraine patients: a systematic review with meta-analysis.
in about 40% of patients with chronic facial pain, even in case of J Headache Pain 2012;13:339e49.
[17] Lipton RB, Dodick DW, Silberstein SD, et al. Single-pulse transcranial magnetic
failure of drug treatment or interventional procedures. stimulation for acute treatment of migraine with aura: a randomised, double-
One of the principles for the therapeutic use of rTMS is that its blind, parallel-group, sham-controlled trial. Lancet Neurol 2010;9:373e80.
effects on synaptic plasticity persist beyond the time of stimulation. [18] Brighina F, Piazza A, Vitello G, et al. rTMS of the prefrontal cortex in the
treatment of chronic migraine: a pilot study. J Neurol Sci 2004;227:67e71.
In the literature, motor cortex rTMS delivered at high-frequency [19] Conforto AB, Amaro Jr E, Gonçalves AL, et al. Randomized, proof-of-principle
5 Hz relieves pain by more than 30% in 46%e62% of treated pa- clinical trial of active transcranial magnetic stimulation in chronic migraine.
tients with chronic neuropathic pain [35,36], but reported results Cephalalgia 2014;34:464e72.
[20] Misra UK, Kalita J, Bhoi SK. High-rate repetitive transcranial magnetic stimu-
are usually of short duration. So, the repetition of rTMS sessions is
lation in migraine prophylaxis: a randomized, placebo-controlled study.
essential for producing lasting relief by a cumulative effect. As J Neurol 2013;260:2793e801.
aforementioned, in two studies of patients with facial pain [28,29], [21] Zaghi S, DaSilva AF, Acar M, Lopes M, Fregni F. One-year rTMS treatment for
a daily session of motor cortex rTMS for 5 days provided significant refractory trigeminal neuralgia. J Pain Symptom Manage 2009;38:e1e5.
[22] Hirayama A, Saitoh Y, Kishima H, et al. Reduction of intractable deaf-
pain relief lasting for at least two weeks after the last session. Long- ferentation pain by navigation-guided repetitive transcranial magnetic stim-
lasting analgesic effect of motor or prefrontal cortex rTMS was also ulation of the primary motor cortex. Pain 2006;122:22e7.
 H. Hodaj et al. / Brain Stimulation 8 (2015) 801e807 807

[23] André-Obadia N, Mertens P, Gueguen A, Peyron R, Garcia-Larrea L. Pain relief [41] Agostoni E, Frigerio R, Santoro P. Atypical facial pain: clinical considerations
by rTMS: differential effect of current flow but no specific action on pain and differential diagnosis. Neurol Sci 2005;26(Suppl. 2):s71e4.
subtypes. Neurology 2008;71:833e40. [42] Rossini PM, Barker AT, Berardelli A, et al. Non-invasive electrical and magnetic
[24] Lefaucheur JP, Drouot X, Nguyen JP. Interventional neurophysiology for pain stimulation of the brain, spinal cord and roots: basic principles and pro-
control: duration of pain relief following repetitive transcranial magnetic cedures for routine clinical application. Report of an IFCN committee. Elec-
stimulation of the motor cortex. Neurophysiol Clin 2001;31:247e52. troencephalogr Clin Neurophysiol 1994;91:79e92.
[25] Lefaucheur JP, Drouot X, Menard-Lefaucheur I, et al. Neurogenic pain relief by [43] Busner J, Targum SD. The clinical global impressions scale: applying a research
repetitive transcranial magnetic cortical stimulation depends on the origin tool in clinical practice. Psychiatry (Edgmont) 2007;4:28e37.
and the site of pain. J Neurol Neurosurg Psychiatry 2004;75:612e6. [44] Mhalla A, Baudic S, Ciampi de Andrade D, et al. Long-term maintenance of the
[26] Lefaucheur JP, Hatem S, Nineb A, et al. Somatotopic organization of the analgesic effects of transcranial magnetic stimulation in fibromyalgia. Pain
analgesic effects of motor cortex rTMS in neuropathic pain. Neurology 2011;152:1478e85.
2006;67:1998e2004. [45] Coffey RJ, Lozano AM. Neurostimulation for chronic noncancer pain: an
[27] Lefaucheur JP, Drouot X, Ménard-Lefaucheur I, Keravel Y, Nguyen JP. Motor evaluation of the clinical evidence and recommendations for future trial de-
cortex rTMS in chronic neuropathic pain: pain relief is associated with ther- signs. J Neurosurg 2006;105:175e89.
mal sensory perception improvement. J Neurol Neurosurg Psychiatry [46] Rossi S, Hallett M, Rossini PM, Pascual-Leone A, Safety of TMS Consensus
2008;79:1044e9. Group. Safety, ethical considerations, and application guidelines for the use of
[28] Khedr EM, Kotb H, Kamel NF, Ahmed MA, Sadek R, Rothwell JC. Longlasting transcranial magnetic stimulation in clinical practice and research. Clin Neu-
antalgic effects of daily sessions of repetitive transcranial magnetic stimula- rophysiol 2009;120:2008e39.
tion in central and peripheral neuropathic pain. J Neurol Neurosurg Psychiatry [47] Hadley D, Anderson BS, Borckardt JJ, et al. Safety, tolerability, and effectiveness
2005;76:833e8. of high doses of adjunctive daily left prefrontal repetitive transcranial mag-
[29] Fricová J, Klírová M, Masopust V, Novák T, Vérebová K, Rokyta R. Repetitive netic stimulation for treatment-resistant depression in a clinical setting. J ECT
transcranial magnetic stimulation in the treatment of chronic orofacial pain. 2011;27:18e25.
Physiol Res 2013;62(Suppl. 1):S125e34. [48] Short EB, Borckardt JJ, Anderson BS, et al. Ten sessions of adjunctive left
[30] Nguyen JP, Nizard J, Keravel Y, Lefaucheur JP. Invasive brain stimulation for prefrontal rTMS significantly reduces fibromyalgia pain: a randomized,
the treatment of neuropathic pain. Nat Rev Neurol 2011;7:699e709. controlled pilot study. Pain 2011;152:2477e84.
[31] Pagano RL, Fonoff ET, Dale CS, Ballester G, Teixeira MJ, Britto LR. [49] Herwig U, Fallgatter AJ, Höppner J, et al. Antidepressant effects of augmen-
Motor cortex stimulation inhibits thalamic sensory neurons and enhances tative transcranial magnetic stimulation: randomised multicentre trial. Br J
activity of PAG neurons: possible pathways for antinociception. Pain Psychiatry 2007;191:441e8.
2012;153:2359e69. [50] O’Reardon JP, Solvason HB, Janicak PG, et al. Efficacy and safety of transcranial
[32] Maarrawi J, Peyron R, Mertens P, et al. Motor cortex stimulation for pain magnetic stimulation in the acute treatment of major depression: a multisite
control induces changes in the endogenous opioid system. Neurology randomized controlled trial. Biol Psychiatry 2007;62:1208e16.
2007;69:827e34. [51] Sgambato V, Abo V, Rogard M, Besson MJ, Deniau JM. Effect of electrical
[33] de Andrade DC, Mhalla A, Adam F, Texeira MJ, Bouhassira D. Neuropharma- stimulation of the cerebral cortex on the expression of the Fos protein in the
cological basis of rTMS-induced analgesia: the role of endogenous opioids. basal ganglia. Neuroscience 1997;81:93e112.
Pain 2011;152:320e6. [52] Sgambato V, Pagès C, Rogard M, Besson MJ, Caboche J. Extracellular signal-
[34] Lefaucheur JP, Drouot X, Ménard-Lefaucheur I, Keravel Y, Nguyen JP. Motor regulated kinase (ERK) controls immediate early gene induction on cortico-
cortex rTMS restores defective intracortical inhibition in chronic neuropathic striatal stimulation. J Neurosci 1998;18:8814e25.
pain. Neurology 2006;67:1568e74. [53] Lefaucheur JP. Why image-guided navigation becomes essential in the prac-
[35] Lefaucheur JP, André-Obadia N, Poulet E, et al. Recommandations françaises tice of transcranial magnetic stimulation. Neurophysiol Clin 2010;40:1e5.
sur l’utilisation de la stimulation magnétique transcrânienne répétitive [54] Pedersen JL, Barloese M, Jensen RH. Neurostimulation in cluster headache: a
(rTMS): règles de sécurité et indications thérapeutiques. Neurophysiol Clin review of current progress. Cephalalgia 2013;33:1179e93.
2011;41:221e95. [55] May A, Leone M, Boecker H, et al. Hypothalamic deep brain stimulation in
[36] Lefaucheur JP, André-Obadia N, Antal A, et al. Evidence-based guidelines on positron emission tomography. J Neurosci 2006;26:3589e93.
the therapeutic use of repetitive transcranial magnetic stimulation (rTMS). [56] Magis D, Bruno MA, Fumal A, et al. Central modulation in cluster headache
Clin Neurophysiol 2014;125:2150e206. patients treated with occipital nerve stimulation: an FDG-PET study. BMC
[37] O’Connell NE, Wand BM, Marston L, Spencer S, Desouza LH. Non-invasive Neurol 2011;11:25.
brain stimulation techniques for chronic pain. Cochrane Database Syst Rev [57] Nesbitt AD, Goadsby PJ. Cluster headache. BMJ 2012;344:e2407.
2014;4. CD008208. [58] Lefaucheur JP, Ménard-Lefaucheur I, Goujon C, Keravel Y, Nguyen JP. Predic-
[38] Headache Classification Committee of the International Headache Society tive value of rTMS in the identification of responders to epidural motor cortex
(IHS). The international classification of headache disorders, 3rd edition (beta stimulation therapy for pain. J Pain 2011;12:1102e11.
version). Cephalalgia 2013;33:629e808. [59] Passard A, Attal N, Benadhira R, et al. Effects of unilateral repetitive trans-
[39] Headache Classification Committee of the International Headache Society. cranial magnetic stimulation of the motor cortex on chronic widespread pain
Classification and diagnostic criteria for headache disorders, cranial neuralgias in fibromyalgia. Brain 2007;130:2661e70.
and facial pain. Cephalalgia 1988;8(Suppl. 7):1e96. [60] Lefaucheur JP, Drouot X, Ménard-Lefaucheur I, Nguyen JP. Neuropathic pain
[40] ICD-10: International statistical classification of diseases and related health controlled for more than a year by monthly sessions of repetitive trans-
problems. 2nd ed. Geneva: World Health Organization Editions; 2004. 10th cranial magnetic stimulation of the motor cortex. Neurophysiol Clin 2004;
revision. 34:91e5.