9.9 Neurostimulation in pain management

Pain is highly prevalent and exerts a huge toll on individuals, families and society. A 2011 review conducted by the US Institute of Medicine exemplifies the scope of the problem. It found that 100 million American adults have chronic pain and estimated the annual cost burden to society at US$560–635 billion plus US$100 billion spent by the federal government (Institute of Medicine of the National Academies, 2011). Pain prevalence is likely to rise in most countries as age-related chronic illnesses increase and the long-term effects of HIV/AIDS evolve. Both health systems and individual health professionals face enormous challenges in providing safe and effective care for diverse populations with chronic pain.

Pain is a major factor in the burden associated with serious or life-threatening illnesses, and both the prevalence and the adverse consequences increase in advanced illness. Most studies have focused on populations with cancer and have determined that the prevalence of pain associated with varied solid tumours is 39–50% (Deandrea et al., 2014). The adverse effects of unrelieved pain on physical functioning, mood, coping and adaptation, and caregiver distress justify the widely-held view that pain management is a moral imperative in the clinical management of patients with advanced illness.

There is a strong international consensus that the first-line treatment for pain associated with serious or life-threatening illnesses is opioid therapy. Although broad experience with opioid therapy indicates that it is usually safe and effective, patients may not be able to access optimal therapy or may be unable to obtain satisfactory relief even when treatment is available. Indeed, some studies indicate a relatively high rate of inadequate pain relief (Nekolaichuk et al., 2013), the reasons for which are likely to be multifactorial and variable across settings.

Both non-opioid pharmacological therapies and non-pharmacological interventions are needed to optimize analgesic outcomes. If opioid therapy does not yield a favourable balance between analgesia and side effects, the patient should be considered to be poorly responsive to the current regimen and an alternative strategy should be selected. There are many options. Some, such as opioid rotation, the use of non-opioid or adjuvant analgesics, and varied interventions like neural blockade and neuraxial analgesia, are considered routinely.

Neurostimulation is another important strategy for pain. The term includes an array of interventions that involve precisely targeted stimulation of peripheral nerve, spinal cord, or the brain (Table 9.9.1). Some of these treatments, such as transcutaneous electrical nerve stimulation (TENS) and spinal cord stimulation (SCS), have been available for decades. Others, such as peripheral nerve stimulation (PNS) and transcranial stimulation techniques, have emerged more recently as viable options in pain management. Neurostimulation techniques are seldom used in the management of pain related to serious illness. A better understanding of the available treatments and the emergence of newer technologies, may increase access and use in the future.

The treatments categorized as neurostimulation techniques are highly variable. They target different tissues with different stimulation modalities and use varied technologies to accomplish intended effects.

TENS is a non-invasive technique that uses electrodes on the skin to deliver electrical stimulation to peripheral nerves. The site selected for stimulation usually is in the region of the painful site, but may be at a distance, usually along the course of the peripheral nerves innervating the site (Atamaz et al., 2012). The approach is sometimes considered a rehabilitative modality comparable to heat, cold, vibration, or ultrasound; when used in this way, it may be directed by physical therapists as part of a programme intended to reduce symptoms and improve function. In some countries, however, TENS units may be acquired by patients for home use and applied primarily for the purpose of pain management.

TENS was developed after publication of the gate control theory, which suggested that stimulation of large-diameter, non-nociceptive primary afferent nerves may be able to modulate pain due to interactions with nociceptive pathways in the spinal cord (Johnson et al., 1991; Sluka and Walsh, 2003). Later studies demonstrated that stimulation of nerves can produce dose-dependent (80 Hz vs 30 Hz) segmental inhibition of pressure-pain thresholds mediated by second-order neurons in the dorsal horn of the spinal cord (Chen and Johnson, 2009). Adenosine pathways are responsible, at least in part, for the TENS analgesia, as caffeine and adenosine antagonists can block the effect (Marchand and Charest, 1995). In addition, the endogenous opioid peptides, encephalins and dynorphins, may play a role and may actually mediate differential responses to low versus high TENS frequencies. Specifically, low-frequency TENS stimulation causes met-encephalin release (Han et al., 1991), implicating mu opioid receptors, and high-frequency TENS induces dynorphin A release that activates delta receptors (Kalra et al., 2001) at supraspinal and spinal levels (Sluka et al., 1999; Leonard et al., 2010).

The extent to which the effects of TENS is influenced by the placement of electrodes in the dermatome affected by injury has suggested the possibility that the effects of TENS in the spinal cord may be augmented or modified by more direct actions on peripheral nerves (Walsh et al., 1998). TENS can ameliorate chronic pain when applied to a region innervated by a specific peripheral nerve territory (Gersh et al., 1980; Engholm and Leffler, 2010) and TENS application to dermatomes can increases local pain threshold (Bjordal et al., 2007) and reduce postoperative pain (Chen et al., 1998; Yeh et al., 2010). The potential efficacy of TENS placed at traditional acupuncture points also suggests the importance of peripheral localization of the electrical stimulation (Ng and Hui-Chan, 2007, 2009).

The clinical use of TENS is largely based on favourable anecdotal experience. There have been few high-quality studies capable of determining safety and efficacy in different patient populations. Of the 43 studies in a recent systematic review of TENS treatment for cancer pain, only two were randomized controlled trials (RCTs), 16 were non-randomized, 19 were educational articles, five were on non-cancer related pain, and one was on acute pain (Robb et al., 2009).

The two RCTs were too small to account for variability in settings and patients characteristics, and were considered yield inconclusive results. One of these studies randomly assigned 49 breast cancer patients with thoracic pain due to mastectomy to receive TENS, sham TENS, or transcutaneous spinal electro-analgesia (TSE) (Robb et al., 2007). Treatment for 3 weeks at the intensity recommended by the manufacturers produced no difference among the three arms in terms of the primary outcomes (pain intensity, anxiety, mood, and function); there was, however, less pain interference in the patients randomized to the TENS arm, and at study end, twice as many TENS than TSE patients decided to continue treatment. The second RCT randomized 15 cancer patients with advanced illness to sham, acupuncture-like TENS, or no treatment and found that 30-minute treatment periods for 5 consecutive days produced no difference in pain intensity among the groups (Gadsby et al., 1997).

A systematic review of TENS in patients with advanced illness who were receiving palliative care also noted that conclusions about efficacy are not possible given the limited data available (Pan et al., 2000). Most published experience is in the form of case series (Loh and Gulati 2013). More research is needed to determine whether TENS is efficacious and safe for pain associated with cancer or other serious illnesses.

The ability of TENS to increase blood flow at the site of high-frequency stimulation has raised concerns about the safety of TENS in patients with pain caused by tumour masses. Theoretically, the application of TENS near a tumour could increase angiogenesis or tumour spread. Although recent studies indicate that the elevated blood flow at the TENS stimulation site is the result of increased muscle activity, and that high-frequency or low-frequency TENS below motor threshold has no clinically significant effect on blood flow in tumours (Cooperman et al., 1975; Cata et al., 2004), the limited data on safety may justify caution in the use of high-frequency TENS in those with metastatic disease.

PNS for the management of pain is a strategy that became available in 1970s but has received more attention since the 1990s when more sophisticated leads and stimulators were developed (Verrills et al., 2009). Patients receiving PNS undergo implantation of one or more percutaneous leads adjacent to the peripheral nerves innervating a painful region (Long et al., 1981; Hassenbusch et al., 1996; Johnson and Burchiel, 2004). Nerve dissection is not required and the risk of nerve injury is minimal. Although the literature in support of this approach is mostly characterized by retrospective reviews and case series, the potential for this technology may be significant.

PNS has been utilized for a variety of pain syndromes, including occipital neuralgia, headache, and regional pains in the abdominal, pelvic, axial low back, and cervical regions. In a typical report, PNS was used to address pain in the distribution of sensory nerves that was part of a failed back surgery syndrome; all six patients experienced improvement in pain scores, decreased the utilization of pain medications, and experienced an improvement in function (Paicius et al., 2007). There have been no confirmatory controlled trials and no published experience in populations with serious medical illnesses.

Recently, a variation of PNS has been introduced, which is known as ‘PNS cross- talk’ (PNSCT). By creating an electrical circuit between leads, a larger area can be covered than would be possible using conventional PNS (Burgher et al., 2012). PNSCT has been proposed as an alternative to treat chronic regional and axial pain in patients refractory to conventional therapy (Falco et al., 2009). Controlled trials are needed to evaluate whether this approach is better than PNS and could be applied to the treatment of chronic pain in the medically ill.

SCS is achieved by applying an electrical current to a specific area of the spinal cord (Shealy et al., 1967). Stimulation is accomplished using electrodes placed in the epidural space that deliver a current to the underlying posterior columns. Patient who are candidates for this therapy typically undergo an initial trial using percutaneous leads to determine whether stimulation can be provided to the painful area, and if so, whether it is associated with meaningful analgesia (Ghoname et al., 1999). Those who report benefit are offered subcutaneous implantation of the entire system, which includes the electrodes, connecting wires, and a radiofrequency transmitter. The system is equipped with an alarm that alerts when the batteries need to be replaced; newer systems are rechargeable.

The mode of action of SCS has not been elucidated, but like TENS and PNS, the approach was developed based on predictions developed from the gate control theory (Melzack and Wall, 1965). Specifically, stimulation of large-diameter afferents in the posterior columns may be able to segmentally inhibit transmission of impulses originating in small diameter nociceptive afferents. Ongoing research also suggests that SCS may inhibit transmission in the spinothalamic tract through the activation of central descending inhibitory mechanisms, influencing sympathetic efferent neurons, and releasing various inhibitory neurotransmitters (Oakley and Prager, 2002).

Stancak et al. tested patients with failed back surgery syndrome to determine the area of the central nervous system (CNS) where SCS has its effect. Functional magnetic resonance imaging (fMRI) technology showed that during SCS, there was a signal increase in the somatosensory cortex, primary motor cortex, and insula of the leg area and vicinity, while there was a concomitant signal decrease in the primary motor cortex corresponding to the shoulder. Applying a standardized suprathreshold heat stimulus to the lower leg of study subjects produced increases in the secondary somatosensory cortex, insula, thalamus, and cingulate cortex. During simultaneous spinal cord and painful heat stimulations, the left and right temporal poles and the ipsilateral cerebellar cortex were activated more strongly compared with the sum of the activations of the separate stimulations, suggesting modulation of pain-related activation by ongoing SCS.

In another study of SCS mechanisms, Nihashi and colleagues utilized fluorodeoxyglucose (FDG) positron emission tomography (PET) scanning to study metabolic/glucose uptake information in seven patients with complex regional pain syndrome (CRPS) and 13 controls (Nihashi et al., 2004). They found an increase of FDG metabolism in the left thalamus, secondary somatosensory cortex, anterior cingulate cortex, bilateral insula, dorsolateral prefrontal cortex, and bilateral superior temporal gyrus in the six patients where SCS was effective (defined by > 50% pain reduction). However, FDG uptake decreased in the posterior cingulate cortex.

Most reports of SCS have described case series of patients with neuropathic pains, such as radiculopathies, arachnoiditis (De La Porte and Siegfried, 1983), phantom limb pain, CRPS (Taylor 2006; Taylor et al., 2006), deafferentation syndromes (Sanchez-Ledesma et al., 1989), non-specific neuropathic pain (Kumar et al., 2008), brachial and lumbosacral plexopathies, and post-herpetic neuralgia (Tasker, 1998). Positive effects also were reported in 18 patients with visceral syndromes, such as interstitial cystitis, chronic abdominal pain, chronic pancreatitis, mediastinal pain (Guttman et al., 2009), and intractable angina (Kemler et al., 2000).

A systematic review of the literature on the use of SCS in cancer patients identified only four trials, all of them small observational or retrospective studies (Lihua et al., 2013). In one study, three of 11 patients reported greater than 50% analgesia and underwent implantation (Meglio et al., 1989). Another study of 14 lung cancer patients who underwent SCS implantation for the treatment of chest pain after surgical or radiological interventions noted a mean 71% decline in opioid use after implantation and a reduction in mean visual analogue score from 7.43 at baseline to 3.07 after 1 month and 2.07 after 12 months (Yakovlev et al., 2010). A survey of 15 patients with intractable, cancer-related, low back pain observed similar benefits after SCS implantation (Yakovlev and Resch, 2011).

A retrospective study of 454 patients undergoing SCS implantation for a variety of pain diagnoses observed that 71% of the patients reported partial to complete pain relief, 58% decreased analgesic use, and 11% stopped using analgesics completely (Shimoji et al., 1993). Fifty-two of these patients had cancer-related pain, and 45 also experienced greater than 50% pain reduction. These retrospective data, like the case series, cannot confirm efficacy or establish the benefits of SCS relative to other treatments used for refractory pain. Nonetheless, the observational data are promising and suggest that a trial of SCS should be considered in patients who have pain related to serious illness that has not responded to conventional pharmacological management.

Widespread use of neuroimaging techniques such as fMRI has shown that patients with chronic pain syndromes develop changes in the brain. For example, imaging shows changes in the excitability and/or somatotopic organization within specific cortical and subcortical regions, including structures generally considered to be part of a pain processing network or ‘pain matrix’ (e.g. areas of the motor and somatosensory cortices and parts of the thalamus) (Cohen et al., 1991; Elbert et al., 1994; Flor 2003). The observation that modulation of these neuroplastic changes may be accompanied by pain relief (Pleger et al., 2005; Birbaumer et al., 1997) has provided a physiological justification for trials of CNS neurostimulation for the treatment of chronic pain.

Both non-invasive and invasive techniques of neuromodulation have been studied in different types of chronic pain syndromes. Results have been promising and suggest that CNS neurostimulation could play a role in the treatment of pain related to serious medical illness.

Transcranial direct current stimulation (tDCS) is a non-invasive technique that applies low-intensity direct current to the scalp (Nitsche and Paulus, 2000). The current penetrates the brain and modulates neuronal excitability by altering the firing rate of individual neurons. There are two modalities depending on current direction: anodal (excitatory) tDCS increases cortical excitability and cathodal (inhibitory) tDCS decreases excitability (Nitsche and Paulus, 2001).

Studies have shown that the analgesic effect of tDCS may be elicited by the application of anodal tDCS over the primary motor cortex Fregni et al., (2006a, 2006b) or cathodal tDCS over the somatosensory cortex (Antal et al., 2008, Knotkova et al., 2009). Two randomized, sham-controlled studies showed that anodal stimulation improved pain in patients with fibromyalgia (Fregni et al., 2006b; Roizenblatt et al., 2007), and another sham-controlled trial showed promising results of tDCS for the treatment of central pain in patients with traumatic spinal cord injury (Fregni et al., 2006a). A case study suggested that cathodal tDCS over the somatosensory cortex also may be able to relieve chronic neuropathic pain (Knotkova et al., 2009).

Two recent sham-controlled studies evaluated the analgesic efficacy of tDCS for the treatment of chronic migraine headache: Antal et al. found that cathodal tDCS applied to the visual cortex resulted in a significant reduction in pain intensity when compared with the sham group (Antal et al., 2011), and Dasilva et al. showed that anodal tDCS applied over the primary motor cortex resulted in a delayed reduction in pain intensity four months after stimulation (Dasilva et al., 2012). Furthermore, tDCS was successfully applied in two patients with trigeminal neuralgia and neuropathic facial pain due to surgical disturbance of the trigeminal nerve, resulting in greater than 50% reduction in pain scores and decreased use of pain medications (Knotkova 2009). A small observational study suggests that tDCS may be useful for the treatment of cancer pain (Silva et al., 2007).

Both studies and clinical experience with tDCS suggest that the analgesic properties are cumulative and that the duration of the analgesic effect outlasts the stimulation but is transitory. Pain usually returns to prestimulation levels. Independent investigators have shown greater pain relief with tDCS over 5 consecutive days as compared to a single session, and demonstrated a time course of the analgesic response that can outlast the last stimulation session by over 9 weeks Fregni et al., (2006a, 2006b; Roizenblatt et al., 2007). However, the duration of the effect varies greatly from patient to patient, and predictions on the magnitude of the response and duration of the effect cannot be made based solely on the underlying pain syndrome, concomitant medications (opioids or adjuvants), co-morbidities like depression or anxiety, or other patient characteristics (e.g. age and gender). Indeed, a recent negative study in a group of patients with painful spinal cord injury suggested that time since the injury may affect the magnitude of the analgesic response (Wrigley et al., 2013).

TDCS may yield benefits for symptoms other than pain. One study showed that tDCS treatment in patients with fibromyalgia increased sleep efficiency by 11.8% and decreased arousal by 35% (Roizenblatt et al., 2007). Others have shown improvement in quality of life and a decrease in the use of pain medications Fregni et al., (2006a, 2006b).

In summary, tDCS has been shown to be effective in relieving pain in a variety of chronic pain syndromes, with potential secondary benefits in sleep, quality of life, and use in pain medications. It is relatively inexpensive, easy to implement, and can be portable so that patients can be educated to self-administer the stimulation at home. These characteristics suggest that tDCS could play a useful role in the treatment of pain in populations with serious medical illness. Studies are needed to confirm these positive outcomes in medically ill populations.

Like tDCS, transcranial magnetic stimulation (TMS) is a non-invasive technique applied directly on the scalp. This technology entails the generation of a powerful magnetic field perpendicular to the brain cortex. This field induces the formation of electrical currents that are parallel to the cortex (similar to tDCS) and can modulate the neuronal excitability in underlying brain structures. The electrical field is relatively restricted in space due to the conical shape of the magnetic field, and as a result, TMS can produce more selective regional neuromodulation than tDCS.

TMS can be delivered as a single stimulation (sTMS) or a train of repetitive stimulations (rTMS). It was first explored for the treatment of pain in 1995 and numerous studies since then have been conducted to test various parameters of frequency, duration and width of the wave. The range of frequencies that have been tested is wide and is divided into high- and low-frequency stimulation based on efficacy.

In general, studies involving patients with various types of neuropathic pain suggest that high-frequency rTMS is more effective in eliciting long-lasting pain relief than low-frequency rTMS. For instance, André-Obadia et al. performed a double-blind study comparing the analgesic effect of a single session of 1 Hz and 20 Hz rTMS with sham stimulation in 14 patients with treatment-resistant neuropathic pain, including trigeminal, central post-stroke, and peripheral brachial plexus neuropathic pain. Although pain relief was achieved immediately after stimulation with both, the analgesic effect was maintained 1 week after stimulation in the 20 Hz group only (André-Obadia et al., 2006). Another sham-controlled study of 48 patients with trigeminal neuralgia or post-stroke pain syndrome reported that 20 Hz rTMS yielded significantly greater improvement in pain after five consecutive daily sessions of stimulation; the positive effect persisted for 2 weeks after treatment, suggesting that repeated rTMS sessions also can result in long-lasting pain relief (Khedr et al., 2005).

Other studies suggest that the outcome of TMS is influenced by the cause and location of the pain, and the somatotropic area where stimulation is applied. Lefaucher et al. conducted a study in 60 patients with intractable pain secondary to a variety of causes, including thalamic stroke, brainstem stroke, spinal cord lesion, brachial plexus lesion, and trigeminal nerve lesion. The overall pain reduction was significantly greater in rTMS as compared to sham stimulation, but rTMS was less effective in patients with brainstem stroke than with other lesions. The best results were obtained in patients with facial pain, rather than those with pain in the upper or lower limbs (Lefaucheur et al., 2004). In a later study, the same authors evaluated the relationship between the cortical stimulation site and pain site in 36 patients with chronic neuropathic pain located on the face or the hand. Interestingly, patients with facial pain experienced significantly better analgesic effects after stimulation to the hand rather than to the face. Similarly, patients with hand pain had greater pain relief after stimulation to the face rather than to the hand. The authors concluded that rTMS was the most effective for pain relief when stimulation was applied to an area adjacent to the cortical representation of the painful zone rather than to the motor cortical area corresponding to the painful zone itself (Lefaucheur et al., 2006).

Recent research in volunteers also has suggested that targets for rTMS treatment of pain may include brain regions that process pain perception, such as prefrontal cortex (Martin et al., 2013).

Some studies suggest that there are limitations in the use of rTMS in pain management. A randomized double-blind, sham-controlled, crossover study of patients with painful diabetic polyneuropathy demonstrated only a transient and modest improvement in pain scores following ten daily 5 Hz rTMS (500 pulses/session) of primary motor cortex (M1) (Hosomi et al., 2013). However, it has been proposed that a modification of the magnet (H-coil) would result in better stimulation of areas that otherwise are not easily accessible, like the lower limbs in diabetic neuropathy, hence achieving better pain relief (Onesti et al., 2013).

TMS may be efficacious in the treatment of headache. One evaluated sTMS as an abortive strategy for migraine and observed that pain relief occurred in 69% of the patients in the sTMS group as compared to 48% in the sham-controlled group (Mohammad et al., 2008). The second did not find a difference between the treatment and placebo groups (Clarke et al., 2006). Lipton et al. conducted a large multicenter, double-blind, sham-controlled study to evaluate the efficacy of a portable, self-administered sTMS device for the treatment of acute migraine, and found that 39% of the patients treated with sTMS were pain-free at 2 hours post-treatment, as compared to 22% of the sham controlled group (Lipton et al., 2010).

Four studies have evaluated the efficacy of rTMS for the treatment of frequent migraines. A small randomized study (Brighina et al., 2004) and an observational study Misra et al., 2012) suggested benefit, but this was not confirmed in two placebo-controlled studies (Teepker et al., 2010; Conforto et al., 2012).

There have been no studies of TMS in populations with acute or chronic pain related to serious illnesses. Extant data suggests that the strategy can be efficacious in some types of neuropathic pain and acute headache. Studies have shown that high-frequency stimulation is more effective in achieving long lasting pain relief than low-frequency stimulation, and that the outcome of TMS is influenced by the origin and location of the pain, and the somatotropic area where stimulation is applied. The best results have been obtained with facial pain, and when applied to an area adjacent to the cortical representation of the painful zone rather than to the motor cortical area corresponding to the painful zone itself. Like tDCS, TMS may prove to be a feasible, non-invasive alternative to pharmacological therapy for the treatment of refractory chronic pain.

Deep brain stimulation (DBS) is an invasive neurostimulatory technique that involves stereotactic implantation of electrodes directly into subcortical areas. It was first used for the treatment of chronic pain in the 1950s by placing electrodes in the hypothalamus (Pool et al., 1956), but later on, other brain areas were also considered to be good targets, including the thalamic nuclei and adjacent structures (Hosobuchi et al., 1973), subthalamic nucleus (Marques et al., 2013), the internal capsule (Adams et al., 1974; Hosobuchi et al., 1975), anterior cingulate cortex (Pereira et al., 2013), and the periventricular and periaqueductal grey area (Richardson and Akil, 1977a, 1977b). The analgesic mechanism of action of DBS for chronic pain is unclear and still debated, but the evidence favours modulation of brain activity (Kringelbach et al., 2007, Montgomery and Baker, 2000; McIntyre et al., 2004) over synaptic depression, synaptic inhibition (Dostrovsky et al., 2000), or depolarization blockade (Beurrier et al., 2001). DBS has been successfully used for the management of facial and head pain such as post-herpetic trigeminal neuralgia (Green et al., 2003), anaesthesia dolorosa (Green and Owen et al., 2006), multiple sclerosis (Hamani et al., 2006), genital pain, brachial plexus injuries, and malignancy (Nandi et al., 2003).

Although the recommendation of DBS for the management of headaches should be made with caution due to the limited evidence of its efficacy, there are reports of successful use of DBS for the management of cluster headaches. Leone et al. published the first case report of DBS in which a single patient with cluster headache underwent stereotactic implantation of an electrode in the posterior ipsilateral hypothalamic grey matter. The headaches stopped after 48 hours of stimulation, and the patient remained pain-free 13 months after implantation (Leone et al., 2001). Since then, multiple studies have been published evaluating the efficacy of DBS for the treatment of cluster headaches, most of which are case series. The largest study was a prospective, open-label study which included 16 patients, 13 of whom reported improvement in pain with DBS, and ten achieved a persistently pain-free state at 23 months’ follow-up (Leone et al., 2006). The underlying mechanism seems to involve down-regulation of the increased regional blood flow in the posterior hypothalamus that occurs during an acute attack, as seen with both PET and fMRI techniques (May et al., 2006).

Kringelbach et al. published a case series comparing DBS with other invasive techniques for the management of chronic pain. Approximately 70% of the 65 patients who underwent DBS in the thalamus and/or periventricular or periaqueductal grey (PVG/PAG) experienced pain relief after 1 week post-implantation, and pain relief persisted more than 1 year in 60% of these patients (Nandi et al., 2002). Another study found very good efficacy of DBS for stroke patients complaining of burning hyperaesthesia (Owen et al., 2006) but less efficacy in patients with stroke overall.

Although the advent of the non-invasive CNS neurostimulation approaches may supplant DBS for pain, rare patients with severe treatment-refractory neuropathic pain still may be considered for a trial of invasive brain stimulation. Patient selection for DBS should be done by an experienced team, including a mental health professional experienced with this type of evaluations (Saint-Cyr and Trepanier, 2000; Lang et al., 2006). Medical contraindications to DBS include ventriculomegaly large enough to impede direct electrode passage to the surgical target (Kringelbach et al., 2010) and uncorrectable coagulopathy. The current recommendation is to utilize DBS as the last resort for patients who do not respond to standard of care.

Although the use of DBS is limited to a highly selected group of patients, the number of DBS implantations is greater than that for other invasive strategies such as motor cortex stimulation (MCS). According to a recent report, at least 1300 patients have been implanted with electrodes for DBS for the treatment of chronic neuropathic pain conditions (Kringelback et al., 2007), while only 400 patients have been implanted with motor cortex stimulators during a comparable period (Brown et al., 2003). The goal is to improve patient selection and thus outcomes. One intriguing possibility is the use of autonomic measures as potential objective markers, as shown by stroke patients’ subjective preference for PVG/PAG stimulation over ventral posterolateral nucleus/ventral posteromedial nucleus, and correlations between analgesic efficacy and cardiovascular effects or burning hyperaesthesia (Green and Wang et al., 2006).

There are several types of complications that can occur as a result of the electrode implantation for DBS, including complications related to surgical procedures, complications due to the implanted devices, and side effects that result from the stimulation itself (Hariz 2002). According to Beric et al., 6.5% of the implantations can result in device-related complications, including infections, electrode fracture or dislocation, and hardware failure (Beric et al., 2001). Infrequent life-threatening complications can also occur and include intracranial haemorrhage (Benabid et al., 1996), haematoma, and paralysis (Beric et al., 2001). In addition, other complications like perioperative haemorrhage, occurred in 2.3% of patients undergoing DBS implantation (Beric et al., 2001). Stimulation-related side effects are the most frequently encountered problems in DBS. The symptom tends to correlate with the anatomical structure being stimulated (Hariz 2002), and can include paraesthesias, dysarthria, dyskinesia, gait disturbances, imbalance, confusion, depression, inappropriate laughter (Krack et al., 2001), and mood or personality changes (Bejjani et al., 2002).

MCS is another invasive neurostimulation method, in which electrodes are surgically implanted in the epidural space to deliver an electric current to the motor cortex. This strategy requires the assistance of neuronavigation, most commonly fMRI or somatosensory evoked potentials, to localize the motor cortex. During the neurosurgical procedure, a lead with four electrodes is positioned above the dura and under the periosteum so that all four contacts are over the precentral gyrus of the motor cortex (Rasche et al., 2006). To verify the positioning of the electrodes, a suprathreshold stimulus is delivered that produces a contralateral motor response in the absence of concomitant sensory sensations (Brown and Barbaro, 2003). After the operation, a trial is performed to determine the best electrode combination and stimulation parameters that elicit the maximum response without significant side effects. A variety of settings have been proposed, but the most commonly utilized are an intensity of 2–3 V (range 0.5–9.5 V), with a frequency of 25–50 Hz (range 15–130 Hz), and a pulse width of 200 microseconds (range 60–450 microseconds). The stimulation can be tailored to the individual patient’s needs by programming alternating cycles of stimulation with periods of no stimulation.

As noted, experience with MCS is far more limited than experience with DBS, and there have been no published trials of MCS in medically ill patients (Nguyen et al., 1998; Cioni and Megglio, 2007; Friedland et al., 2007; Arle and Shils, 2008; Lima and Fregni, 2008). Favourable case reports (Nguyen et al., 2000; Esfahani et al., 2011) are now supplemented by a few clinical trials suggesting positive outcomes from this approach. A randomized trial of MCS in 16 patients with chronic neuropathic pain following peripheral nerve lesions reported that 60% of patients had satisfactory levels of pain relief after 1 year (Lefaucheur et al., 2009).

The potential role of MCS is ill-defined. If non-invasive neurostimulation strategies are available, they will likely be tried first. If an invasive CNS neurostimulation treatment is considered for refractory chronic pain associated with serious illness, the choice between DBS and MCS is likely to revolve around the experience of the clinician.

Stimulation of the peripheral or central nervous systems for the management of chronic pain has been done for many years and the results have been mixed. Over the years, technology has improved and the delivery of current has become more sophisticated, allowing stimulation of discrete areas with a variety of intensities and frequencies that can be tailored to the underlying condition. The most efficacious parameters, stimulation targets, and devices have evolved with the improvement in technology, but selection of the proper stimulation strategy and patient population continues to be based on limited data and clinical experience. Studies are needed to confirm efficacy and illuminate those factors that predict a high likelihood of a favourable response. Further research also may reveal additional innovations, such as the potential utility of combinations of stimulations (e.g. SCS and DBS) (Chodakiewitz et al., 2013).