Key points

Introduction

Chronic pain is a common condition worldwide; affecting approximately one-third of the adult population in the United States. This complex clinical conundrum poses significant management challenges to physicians. In epidemiologic studies, the prevalence of chronic pain is approximately 48% and that of chronic neuropathic pain varies from 6% to 8%. Given its high prevalence in the general population, economists estimate that the annual cost of chronic pain ranges from $560 billion to $635 billion a year in the United States alone. The International Association for the Study of Pain defined neuropathic pain as “pain initiated or caused by a primary lesion or dysfunction in the nervous system.” Neuropathic pain can be further classified as peripheral or central (depending on the site of the disorder) and acute or chronic (lasting >3 months). Based on imaging studies, there is accruing evidence that chronic pain results from alteration in neural networks and central pain mechanisms including perception, sensitization, and pain modulation pathways. Therefore, modulation of these neural networks may provide clinical benefits in patients with chronic pain syndromes.

In 1991, Tsubokawa and colleagues showed the efficacy of motor cortex stimulation in patients with deafferentation pain. Since then, numerous investigators have explored this domain to alleviate symptoms in patients with chronic neuropathic pain. The introduction of noninvasive brain stimulation reduced the morbidity associated with invasive stimulation techniques and fostered interest in this modality for chronic pain syndromes. The noninvasive modalities of neuromodulation include repetitive transcranial magnetic stimulation (rTMS), transcranial direct current stimulation, cranial electrotherapy stimulation, and reduced impedance noninvasive cortical electrostimulation. Of these, rTMS is the most frequently used and studied modality in patients with chronic pain, with variable outcomes. A recent systematic review on the utility of transcranial magnetic stimulation (TMS) in patients with fibromyalgia syndrome concluded that high-frequency rTMS at the left primary motor cortex (M1) significantly reduces the pain associated with fibromyalgia syndrome and has a lasting impact beyond the duration of stimulation. Contrary to this, a recent Cochrane Systematic Review concluded that single doses of rTMS may provide short-term beneficial effects in patients with chronic pain; however, multiple doses of rTMS for chronic pain did not show consistent benefits across different studies. To date, 30 studies have evaluated the efficacy of rTMS in 528 patients with chronic pain syndromes (fibromyalgia, complex regional pain syndrome, phantom limb pain, thalamic pain, poststroke pain, pain related to injury of spinal cord/brachial plexus). This article focuses on the efficacy of rTMS in patients with different pain syndromes and the physiology and technique of rTMS with a review of the pertinent literature.

Introduction

Chronic pain is a common condition worldwide; affecting approximately one-third of the adult population in the United States. This complex clinical conundrum poses significant management challenges to physicians. In epidemiologic studies, the prevalence of chronic pain is approximately 48% and that of chronic neuropathic pain varies from 6% to 8%. Given its high prevalence in the general population, economists estimate that the annual cost of chronic pain ranges from $560 billion to $635 billion a year in the United States alone. The International Association for the Study of Pain defined neuropathic pain as “pain initiated or caused by a primary lesion or dysfunction in the nervous system.” Neuropathic pain can be further classified as peripheral or central (depending on the site of the disorder) and acute or chronic (lasting >3 months). Based on imaging studies, there is accruing evidence that chronic pain results from alteration in neural networks and central pain mechanisms including perception, sensitization, and pain modulation pathways. Therefore, modulation of these neural networks may provide clinical benefits in patients with chronic pain syndromes.

In 1991, Tsubokawa and colleagues showed the efficacy of motor cortex stimulation in patients with deafferentation pain. Since then, numerous investigators have explored this domain to alleviate symptoms in patients with chronic neuropathic pain. The introduction of noninvasive brain stimulation reduced the morbidity associated with invasive stimulation techniques and fostered interest in this modality for chronic pain syndromes. The noninvasive modalities of neuromodulation include repetitive transcranial magnetic stimulation (rTMS), transcranial direct current stimulation, cranial electrotherapy stimulation, and reduced impedance noninvasive cortical electrostimulation. Of these, rTMS is the most frequently used and studied modality in patients with chronic pain, with variable outcomes. A recent systematic review on the utility of transcranial magnetic stimulation (TMS) in patients with fibromyalgia syndrome concluded that high-frequency rTMS at the left primary motor cortex (M1) significantly reduces the pain associated with fibromyalgia syndrome and has a lasting impact beyond the duration of stimulation. Contrary to this, a recent Cochrane Systematic Review concluded that single doses of rTMS may provide short-term beneficial effects in patients with chronic pain; however, multiple doses of rTMS for chronic pain did not show consistent benefits across different studies. To date, 30 studies have evaluated the efficacy of rTMS in 528 patients with chronic pain syndromes (fibromyalgia, complex regional pain syndrome, phantom limb pain, thalamic pain, poststroke pain, pain related to injury of spinal cord/brachial plexus). This article focuses on the efficacy of rTMS in patients with different pain syndromes and the physiology and technique of rTMS with a review of the pertinent literature.

Mechanism of action and parameters of rTMS relevant to chronic pain management

The technique of TMS was introduced in the 1950s, but it was not until the introduction of noninvasive magnetic stimulation by Barker and colleagues in 1985 that the scope of this therapy was widened. Since then TMS has been evaluated for refractory depression, chronic pain, neuropathic pain, schizophrenia, and obsessive compulsive disorders, with varied success. TMS involves generation of action potentials with either activation or inhibition of various cortical and subcortical neural networks by generating a magnetic field that penetrates the skull of the patient. High-frequency rTMS of the M1 has been shown to induce rapid changes and modulation of the sensorimotor networks in healthy individuals. It differs from conventional electrical stimulation in the way currents are induced in the underlying neural tissue. Simple electrical stimulation requires that current flow from an anode to a cathode placed on the scalp activates/inhibits neural tissue below; however, the impedance associated with the scalp and cranium limits the reach of electrical current to the most superficial layers of cortex (without increasing the electrical current to levels that are painful for patients). TMS uses electromagnetic induction to generate an electrical current. A high-current pulse is discharged from a stimulator to a wire coil to produce a magnetic field. When the wire coil is placed tangentially to the scalp, the magnetic field passes unobstructed through the scalp and cranium to excite or inhibit the neural tissue below. Compared with electrical stimulation, magnetic stimulation allows the study of focal neural tissue activation in which the signal is not impeded by other tissues and is minimally invasive to the patient.

As with other types of stimulation, the effect of TMS depends strongly on positioning of the source (ie, position of the coil), the orientation of the underlying cell structure, duration of stimulation, and the stimulation parameters used to activate the underlying neural tissue. The currents induced by TMS flow parallel to the plane of the wire coil. It has been suggested that there is preferential activation of cells that are horizontally positioned within the cortex when the coil placed tangentially to the scalp ( Figs. 1 and 2 ), such that cortical interneurons are preferentially stimulated and cortical pyramidal neurons are activated transsynaptically because of their vertical orientation within the cortex ; however, this may be a generalization because it is difficult to accurately estimate the area of activation because of variation in the cortical geometry, currents generated, cerebral plasticity, membrane potentials, and ion channel activity.

( A ) The positioning of a wire coil (MagVenture) when delivering TMS to motor cortex. The head tracker is adhered to the forehead on the opposite side to where stimulation is to be delivered. The 3 reflecting spheres are used in conjunction with an infrared camera to visualize the target location. ( B ) We use software from Localite to identify stimulation targets. Data from the orientation of the reflective spheres and camera are merged with a T1 magnetic resonance imaging of a patient. When the coil is positioned over the patient’s cranium, the software displays the position of the coils in the sagittal, coronal, and horizontal planes, as well as providing a three-dimensional representation of the patient’s cranium reconstructed from the MRI with the coil position registered in real time.

Simplified diagram of the electrical and magnetic fields generated during TMS. The 2 circular figures represent a butterfly coil configuration for this example ( solid black lines ). Red dashed lines indicate the direction of the flow of current through the wire coils during TMS. Blue dashed lines indicate the direction of the magnetic field generated by the wire coil during stimulation. The black dashed line represents the direction of the currents generated in neural tissue as a result of TMS. The size of the field, penetrating depth through tissue, and level of activation depend on stimulation parameters and coil configuration.

The ability of TMS to modify the synaptic characteristics of neural networks depends on stimulation parameters. The frequency at which the stimulation is delivered largely determines whether the stimulation produces an excitatory or inhibitory effect on the underlying neural tissue. In general, high-frequency stimulation increases the cortical excitability, whereas low-frequency stimulation has inhibitory control on the neural circuits. High-frequency rTMS (>5 Hz) produces an excitatory effect on neural tissue by inducing a form of long-term potentiation (LTP) that increases efficacy at the synapse that can last beyond the duration of a treatment session. High-frequency rTMS delivered to the left M1 has been shown to be effective in the treatment of widespread chronic pain in patients with fibromyalgia, while simultaneously increasing short intracortical inhibition and intracortical facilitation in the motor cortex. Patients with neuropathic pain who had received high-frequency TMS to motor cortex also had reduced intracortical inhibition that correlated with pain relief in the corresponding hand. The ability of high-frequency TMS to increase synaptic efficacy is likely to depend on modulation of both the GABAergic and glutamatergic systems, which have been shown to be important in the modulation of function in motor cortex. Low-frequency rTMS (≤1 Hz) produces an inhibitory effect on neural tissue via a long-term depression (LTD)–like mechanism. However, the excitatory and inhibitory effects of high-frequency and low-frequency rTMS respectively depend on the target and the state of the polarization of the neural tissues being stimulated. Patients with fibromyalgia who received low-frequency rTMS to the right dorsal prefrontal cortex reported improvements in both pain and mood. Low-frequency magnetic stimulation can also be used successfully in the periphery for neuropathic pain management when stimulation is delivered at the site of nerve entrapment. Both high-frequency and low-frequency TMS are effective for pain management and the effectiveness depends on target selection, which suggests that the neural networks involved in pain perception are both widespread and complex.

Theta burst stimulation (TBS) is a form of TMS that also modulates neural function in a way that reflects LTP-like and LTD-like mechanisms that last longer than those induced by rTMS. TBS uses high-frequency bursts (3 pulses at 50 Hz) that are repeated at intervals of 200 milliseconds. Intermittent TBS (iTBS) increases the amplitude of neuron responses and increases cortical excitability in human motor cortex, whereas a continuous TBS (cTBS) pattern suppresses evoked responses. Although the pattern of delivery was thought to determine its excitatory or inhibitory effect on neural tissue, recent work has shown that the effects of iTBS and cTBS may not be as clear. Likewise, the success of TBS for pain management has been mixed. cTBS delivered to the M1 had been reported to reduce laser-evoked pain perception ; however, other studies have not shown either cTBS or iTBS to be effective when applied to M1 alone in patients with neuropathic pain.

Successful pain management with magnetic stimulation may rely on a combination of stimulation types and targets that are tailored to the disease and specific symptoms. For instance, a recent study showed that the analgesic effects of high-frequency TMS are enhanced when iTBS is first delivered to M1, a paradigm known as priming. Although this is an interesting perspective on stimulation-based therapies, more studies are needed to evaluate this approach.

Activation of cortical structures in turn transmits action potentials to various limbic and other pain modulation pathways such as cingulate gyrus, orbitofrontal cortex, insula, hippocampus, caudate nucleus, and periaqueductal areas in either an orthodromic or antidromic direction. rTMS has been shown to induce the release of dopamine in the ipsilateral caudate nucleus, which might have implications in the clinical role of this therapy. rTMS of the M1 and dorsolateral prefrontal cortex has been shown to have antinociceptive and antidepressants effects, respectively. Release of endogenous opioids in the anterior cingulate cortex and periaqueductal gray matter has been implicated in the nociceptive effects of rTMS of M1.

Safety profile of TMS

The excellent safety profile of TMS is the major advantage of this noninvasive treatment modality and led to the expansion of this therapy for a variety of indications. The most serious of the reported complications is the occurrence of seizures, and most of these complications occurred either because of nonadherence to recommended stimulation parameters or in patients who were on medications (tricyclic antidepressants, antipsychotics, theophylline, cocaine, alcohol, amphetamines, MDMA [3,4-methylenedioxy-N-methylamphetamine]) that are known to lower seizure threshold. The risk of epilepsy following TMS is <1% in healthy individuals which increases to 1.4% in patients with history of epilepsy. In addition, TMS in patients with intracranial implants can result in heating or magnetization of these implants and impaired functioning. Therefore TMS is relatively contraindicated in patients with DBS (Deep brain stimulation) electrodes, cochlear implants, or aneurysm clips, whereas it is safer in patients with titanium implants or implants containing ferromagnetic alloys. Other reported adverse effects include histotoxicity, hearing disorders, persistent local pain, transient headache or discomfort at the TMS coil application, impaired cognition, psychiatric symptoms, and biological alterations. In our study, the safety profile of TMS was assessed by noting the occurrence of any of the adverse effects mentioned earlier during the treatment sessions that may or may not have resulted in discontinuation of therapy in either of the groups.

TMS for chronic pain

High-frequency rTMS of the M1 has been shown to induce rapid changes and modulation of the sensorimotor networks in healthy individuals. High-frequency rTMS has also been shown to have a direct effect on sensory thresholds for both cold and hot temperature sensations and thus may be effective in alleviating symptoms in patients with chronic pain. Low-frequency rTMS (1 Hz) over the M1 has been shown to induce early recovery from capsaicin-induced acute pain mediated by C fibers in healthy volunteers compared with sham stimulation and control cohorts. This low-frequency stimulation over M1 was also associated with decreased regional cerebral blood flow (rCBF) in the right medial prefrontal cortex (Brodmann area [BA] 9) and a significant increase in rCBF in the right caudal anterior cingulate cortex (BA 24) and the left premotor area (BA 6). However, the same study group reported facilitation of A-delta fiber–mediated acute pain by low-frequency rTMS over the motor cortex in 13 normal subjects using thulium:yttrium-aluminum-garnet laser stimulation. High-frequency rTMS (10 Hz) of the M1 has similarly not been able to control A-delta fiber–mediated experimental electrically induced acute pain with an increase in the pain scores. In contrast with this, continuous TBS of the M1 has been shown to significantly reduce the perception of laser-induced acute pain in both hands. High-frequency rTMS of the medial (medial frontal cortex) and lateral (motor cortex) pain-regulating pathways has been shown to have variable effects on sensory perception and pain threshold levels. High-frequency stimulation of lateral pain pathways resulted in increase in both sensory perception and pain threshold levels, whereas stimulation of medial pain pathways resulted in decrease in pain threshold levels with an increase in sensory perception levels. There are conflicting reports of TMS of the medial frontal cortex on pain perception, with some of the studies reporting suppression and others facilitating perception of pain following stimulation. There are similarly discordant findings regarding the efficacy of TMS of the sensorimotor cortex on the perception of experimentally induced pain. Low-frequency (1 Hz) rTMS of the right dorsolateral prefrontal cortex (DLPFC) has been shown to increase pain tolerance and threshold to cold pressure test in 180 right-handed healthy volunteers. High-frequency (10 Hz) rTMS over the left prefrontal cortex has been reported to increase thermal pain threshold in 20 healthy adults. Therefore DLPFC can be explored as a potential cortical target to alleviate chronic pain. The results of these studies based on experimentally induced acute pain need to be interpreted cautiously because of the heterogeneity in the rTMS protocol, the cortical target stimulated, and the difference in the pathophysiology of acute and chronic pain mechanisms.

rTMS of M1 for Chronic Pain

In one of the initial studies on the utility of motor cortex stimulation using electrodes in patients with chronic pain, 8 of the 12 patients (67%) benefitted from this therapy after 1 year of stimulation. In 1995, Migita and colleagues reported the utility of rTMS (0.2 Hz using circular coils) applied to the contralateral M1 in 2 patients with central pain. One patient with central pain secondary to left putaminal hemorrhage reported significant pain relief following rTMS of contralateral M1, which was comparable with electrical cortical stimulation. However, another patient with cerebral palsy who underwent left thalamotomy twice at ages 9 and 13 years did not report attenuation of pain following rTMS of motor cortex. Lefaucher and colleagues in 2001 reported significant pain relief in 18 patients with chronic neurogenic pain (thalamic or brainstem stroke, n = 12; brachial plexus lesion, n = 6) following 10-Hz rTMS over the motor cortex compared with 0.5-Hz or sham stimulation. Later, the same group reported transient pain relief for about a week in 14 patients with thalamic stroke (n = 7) and trigeminal nerve lesion pain (n = 7) following a 20-minute session of 10-Hz rTMS over the motor cortex. In 2003, Canavero and colleagues reported that 50% of patients responded to motor cortex stimulation using rTMS (0.2 Hz, figure-of-eight and double-cone coils) in the long term in their series of 9 patients with central pain (poststroke pain n = 5, and spinal cord lesion pain n = 4). Another study using 20-Hz rTMS of the primary cortex in 12 patients with chronic pain (spinal cord lesion n = 2, osteomyelitis n = 1, complex regional pain syndrome n = 2, phantom limb pain n = 1, peripheral nerve lesion n = 6) did not reveal a significant difference between active (mean visual analog scale [VAS] reduction, −4.0% ± 15.6%) and sham stimulation (mean VAS reduction, −2.3% ± 8.8%) in terms of pain relief. Numerous studies have reported the efficacy of rTMS (10 Hz) of the M1 in relieving chronic pain related to thalamic/brainstem stroke, trigeminal nerve, brachial plexus, nerve trunk, and spinal cord lesions compared with sham stimulation. One of the studies reported that rTMS is more effective when applied to an area adjacent to the cortical representation of the painful zone compared with stimulation of the motor cortex corresponding with the painful area (facial pain improved instead of hand pain when the hand motor cortical area was stimulated). Pleger and colleagues reported transient pain relief in 70% of patients (7 of 10 patients) with complex regional pain syndrome following rTMS (10 Hz) of the motor cortex, with maximum benefits being observed 15 minutes after stimulation. Similar results were reported by Picarelli and colleagues in 23 patients with refractory complex regional pain syndrome affecting the upper limbs, and rTMS was recommended as an add-on therapy in such patients. In contrast with this, Irlbacher and colleagues reported no significant difference in pain relief following sham stimulation or rTMS (1 Hz or 5 Hz) in 27 patients with central (n = 13) and phantom limb pain (n = 14). High-frequency (20 Hz) rTMS of the motor cortex has been shown to have long-lasting effects (2 weeks) on pain control in patients with poststroke or trigeminal nerve lesion pain. Low-frequency rTMS (1 Hz) has been shown to have a proalgesic effect, whereas high-frequency rTMS (20 Hz) predicted the efficacy of subsequent motor cortex stimulation in a double-blind study. A randomized double-blind, sham-controlled, crossover study (7 centers, n = 70 patients) concluded that daily multisession (10 sessions) high-frequency (5 Hz) rTMS of M1 provides short-term and modest pain relief in patients with neuropathic pain. Therefore, low-frequency (1 Hz or 5 Hz) rTMS of M1 is unlikely to induce significant beneficial effects in terms of pain control compared with high-frequency (10 Hz or 20 Hz) stimulation ( Table 1 ).

Table 1

Studies showing the efficacy of rTMS in patients with chronic pain syndromes

Study	Origin of Pain (and # Patients)	Target	Stimulation Parameters	Pain Outcome
Migita et al, 1995	Cerebral palsy and thalamotomy (n = 1) and putamen hemorrhage (n = 1)	M1	0.2 Hz, 80% of stimulator output	Successful pain relief in 1 patient
Lefaucheur et al, 2001	Neuropathic pain (n = 18)	M1	0.5 Hz or 10 Hz, 80% RMT	20% reduction in pain symptoms at 10 Hz
Lefaucheur et al, 2001	Thalamic stroke (n = 7) and trigeminal neuropathy (n = 7)	M1	10 Hz, 80% RMT	30% reduction in pain symptoms for up to 8 d
Reid et al, 2001	Teeth removal (n = 1)	Left prefrontal cortex	20 Hz, 100% RMT	42% decrease in pain lasting for 1 mo
Canavero et al, 2003	Stroke (n = 5) and lesion (n = 4)	M1	0.2 Hz, 100% of stimulator output	50% of patients were responsive to treatment
Rollnik et al, 2002	Varying chronic pain syndromes (n = 12)	M1	20 Hz, 80% RMT	Analgesic effects in only some patients
Topper et al, 2003	Root avulsion (n = 2)	Posterior parietal cortex	1 Hz or 10 Hz, 110% RMT	Transient pain relief only
Lefaucheur et al, 2004	Stroke (n = 24) and trigeminal nerve, brachial plexus, and spinal cord lesion (n = 36)	M1	10 Hz, 80% RMT	Pain level reduction in 65% of patients, better results with facial pain
Lefaucheur et al, 2004	Brachial plexus lesion (n = 1)	M1	10 Hz, 80% RMT	Pain controlled with monthly sessions of rTMS for 16 mo
Pleger et al, 2004	Complex regional pain syndrome (n = 10)	M1	10 Hz, 110% RMT	Pain decreased in 70% of patients, but effect was transient
Kheder et al, 2005	Neuropathic pain (n = 48)	M1 hand area	20 Hz, 80% RMT	Pain relief that lasted 2 wk
Fregni et al, 2005	Chronic pancreatitis (n = 5)	Secondary somatosensory cortex	1 Hz or 20 Hz, 90% RMT	36% improvement in pain with 1-Hz, improvement with left-side stimulation only
Andre-Obadia et al, 2006	Stroke (n = 10) and lesion (n = 4)	M1	1 Hz or 20 Hz, 90% RMT	Repetitive 20 Hz most effective for analgesia
Lefaucheur et al, 2006	Chronic neuropathic pain of face and hand, stroke (n = 9) and lesion (n = 27)	M1	10 Hz, 90% RMT	Best efficacy with stimulation to the hand and face area of M1
Hirayama et al, 2006	Deafferentation pain	M1, S1, premotor, SMA	5 Hz, 90% RMT	Pain relief in 50% of patients stimulated in M1 only, relief lasted 3 h
Sampson et al, 2006	Fibromyalgia (n = 4)	DLPFC	1 Hz, 110% RMT	Good analgesic effect that lasts >15 wk
Passard et al, 2007	Fibromyalgia (n = 30)	M1	10 Hz, 80% RMT	Long-lasting decrease in chronic widespread pain
Lefaucheur et al, 2008	Neuropathic pain (n = 48)	M1	1 Hz or 10 Hz, 90% RMT	Improvement in thermal sensory perception to the painful region that correlated with pain relief
Brockardt et al, 2009	Chronic neuropathic pain (n = 4)	Prefrontal cortex	10 Hz, 100% RMT	Decrease of 19% in daily pain
Carretero et al, 2009	Fibromyalgia (n = 28)	DLPFC	1 Hz, 110% RMT	No difference
Lefaucheur et al, 2010	Chronic neuropathic pain in 1 upper limb (n =32)	M1, hand area	10 Hz, 90% RMT	Reduction of laser-induced pain scores in patients with pain
Picarelli et al, 2010	Complex regional pain syndrome type 1 (n = 23)	M1	10 Hz, 100% RMT	Reduction in pain intensity
Mhalla et al, 2011	Fibromyalgia (n = 40)	M1	10 Hz, 80% RMT	Long-term, reduced pain intensity
Short et al, 2011	Fibromyalgia (n = 20)	Prefrontal cortex	10 Hz, 120% RMT	29% reduction in pain symptoms
Lee et al, 2012	Fibromyalgia (n = 15)	DLPFC (1-Hz group), and M1 (10-Hz group)	1 Hz, 110% RMT or 10 Hz, 80% RMT	Low-frequency group (DLPFC) had greater improvement in pain and antidepressant effects relative to high-frequency group (M1)
Hosomi et al, 2013	Neuropathic pain (n = 70)	M1	5 Hz, 90% RMT	Transient, modest pain relief
Tzabazis et al, 2013	Fibromyalgia (n = 16)	Dorsal anterior cingulate cortex	1 Hz or 10 Hz, 110% RMT	43% reduction in pain score at 10 Hz, no effect at 1 Hz

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Introduction to Neuropathic Pain Syndromes Intracranial Ablative Procedures for the Treatment of Chronic Pain Percutaneous Treatments for Trigeminal Neuralgia Microvascular Decompression for Trigeminal Neuralgia Multimodal Treatment of Pain Nerve Blocks for Chronic Pain

Stay updated, free articles. Join our Telegram channel

Join