Peripheral Neurectomy

Known from the sixteenth century, thanks to Ambroise Paré, neurectomy consists of the resection of a portion or the whole of one or more peripheral nerves. It is an easy and simple procedure, performed under local anesthesia, but its long-term failure rate makes its indications in chronic pain management very limited today.¹² In fact, adjacent intact sensory nerves can reinnervate the anesthetic region, thus causing pain recurrence; a new pain sensation due to the denervation hypersensitivity and painful postoperative neuromas may develop, too.^12,13

Neurectomy is indicated in those cases in which pain is limited to the nervous distribution of a specific nerve; such cases are rare and it is thus preferable (given the innervation’s juxtaposition of different dermatomes) to perform peripheral neurotomies in steps. Painful neuromas following peripheral nerve injuries may benefit from neuroma excision and neurectomy, if conservative nonsurgical treatment has not succeeded.^13,14 Proximal neurectomy in amputee patients with painful stump neuromas can be of benefit, but nonspecific stump pains or phantom limb pains do not respond to nerve resection.^12,15 Chest wall cancer pain and postherpetic neuralgia also do not benefit from neurectomy.

In some rare cases it is possible to perform a trial infiltration of long-acting anesthetic, eventually followed by neurolysis with percutaneous thermocoagulation or with injection of neurodestructive agents, such as phenol or alcohol;¹⁶ potential action sites are the posterior ramus of the second cervical nerve in the case of occipital neuralgia and the posterior rami of the spinal intercostal nerves in the case of drug-refractory intercostal neuralgia.

Dorsal Rhizotomy and Dorsal Root Ganglionectomy

Dorsal rhizotomy implies the sectioning of the dorsal nerve roots. Extradural, percutaneous, and radiofrequency methods have been described as well as partial rhizotomy.^17–20 The procedure was indicated especially in the thoracic and sacral regions for malignant pain,²¹ but because of the high rate of pain recurrence^22,23 and development of dysesthesic or deafferentation pain²⁴ the procedure has been all but abandoned these days.

Performed first in 1966 by Scoville,²⁵ based on the discovery that a certain number of axons in the ventral root were unmyelinated afferent fibers from the dorsal root ganglion,^26–28 dorsal root ganglionectomy showed some long-term pain relief in the treatment of chest wall pain, known as the postthoracotomy syndrome.¹¹ Some authors suggest it for sciatica, too, but there is some controversy about this indication.

Dorsal Root Entry Zone Lesions

The dorsal root entry zone (DREZ) is an anatomical entity that includes the central portion of dorsal root, the tract of Lissauer (TL), and layers I to V of the dorsal horn, where the afferent fibers synapse with the origins of the sensory pathways.²⁹ The TL plays an important role in the intersegmental modulation of the nociceptive afferents. It is located dorsolaterally to the dorsal horn and it is divided in two parts: (1) the medial part receives nociceptive afferents directed to the dorsal horn and transmits the excitatory effects of each dorsal root to the adjacent segment;^30,31 (2) the lateral part receives longitudinal endogenous propriospinal fibers interconnecting different levels of substantia gelatinosa (SG) and thus conveying the inhibitory influence of the SG into the neighboring metameres.³¹ The SG, through synaptic connections with the dendrites of some neurons of the spinoreticulothalamic (STR) tract, exerts a strong modulating effect on the nociceptive input (Fig. 49.1A). When the large lemniscal afferents in peripheral nerves or dorsal roots are altered, there is reduction of the inhibitory control of the dorsal horn, hence resulting in excessive firing of the dorsal horn neurons and causing deafferentation pain.^32–35

FIGURE 49.1 A, Transverse hemisection of the spinal cord at the lower cervical level with myelin stained by luxol-fuchsin showing the myelinated rootlet afferents that reach the dorsal column (DC). Small arrow: Pial ring of the dorsal rootlets. Large arrow: Microsurgical DREZotomy target. B, Scheme of DREZ area and DREZotomy target. The pial ring (AP), located about 1 mm before the penetration of the rootlet into the dorsolateral sulcus, is considered the point of transition from the peripheral to the central segment of the rootlet. Peripherally, the fibers are mixed together, but at the level of the AP, the fine nociceptive fibers run more laterally on the rootlet surface. In the central segment, they group in the ventrolateral portion of the DREZ to enter the dorsal horn (DH) through the tract of Lissauer (TL). The lemniscal fibers (Lem) run dorsomedially in the deeper part of the rootlet, and the myotatic fibers (myot) run in the middle of the DREZ. The black triangle indicates the proposed extent of the surgical lesion. C, Intraoperative view during microDREZotomy. While the sucker is elevating the dorsal rootlet, the dissector is pointing at the DREZotomy target. DREZ, dorsal root entry zone.

(A modified from Fig. 56.1 in Burchiel KJ, editor. Surgical Management of Pain. New York: Thieme; 2002.)

Following avulsion injuries of the brachial or lumbosacral plexus between 20% and 90% of patients may develop severe deafferentation pain either immediately after injury or delayed until up to 3 months.

The first DREZ open surgery procedure was performed by Sindou in 1972 on a patient with chronic pain due to Pancoast syndrome.³⁶ Soon after him, in 1974, Nashold started using radiofrequency thermocoagulation.^37,38 Later, DREZ procedures were performed with laser^39,40 and with the use of ultrasound probes.^41,42

The procedure known as microsurgical DREZotomy (MDT) was again accomplished by Sindou²⁹ and involves several segments of the lateral portion of the dorsal horn (where C fibers predominate), meaning the destruction of the LT and laminae I to V of Rexed (Fig. 49.1B and C). By sparing the medial portion of the dorsal horn where large A fibers enter, Sindou tried to avoid complete abolition of tactile and proprioceptive sensation, preserving antinociception and preventing deafferentation.⁴³ Intraoperative sensory evoked potentials (SEPs) during MDT are recommended for identification of the spinal cord segments and can also be helpful in monitoring the surgical procedure itself.^44–46 Sindou developed special microelectrodes for intraoperative microelectrophysiology studies and performed microdialysis studies in the dorsal horn as well.^47,48

MDT has several indications. Some cancer pain patients may benefit from MDT. Good candidates are patients with long life expectancy, sufficient general conditions to undergo surgery and general anesthesia, and topographically delimited pain due to well-localized tumors. MDT from C7 to T2 can be performed for Pancoast syndrome.⁴⁹ Malignancies with circumscribed invasion of the thorax, of the abdomen wall, of the perineal floor, or lumbosacral roots or plexus are other indications for MDT. For more extensive cancer, intrathecal opioids or high cervical anterolateral cordotomy is preferable.²⁹

Patients with chronic deafferentation pain from plexus avulsion are tremendously difficult to manage for pain doctors. Even if in some cases initial conservative therapy may be of some help, standard analgesics and narcotics are usually ineffective and medical management is limited to antidepressant, sedative, and anticonvulsant medications combined with psychological support.⁵⁰ DREZ lesion surgery that is not limited to the avulsed segments but is extended to the adjacent roots is instead recommended for plexus avulsion pain. The results are extremely good, with more than 85% of patients reporting good to excellent pain control after brachial avulsion,^6,36,51–54 although pain relief is fairly lower for conus medullaris root avulsion.⁵⁵ Mortality rate is low,^6,51 and morbidity consists of transient sensory deficit and mild weakness.⁶ DREZ lesion surgery for lumbosacral avulsion carries less than 10% risks of bladder and sexual dysfunction.⁵⁵

Pain following spinal cord injuries (SCIs) can have a multitude of causes, and hence rigorous patient assessment is mandatory. MDT is indicated for burning pain with radicular distribution at the level of the lesion or for patients complaining mainly of allodynia and electric shock–like border-zone pain. Conversely, pain in the totally or almost totally anesthetic area below the lesion, especially in the perineosacral region, is not favorably influenced by MDT.²⁹ These patients should be referred to rehabilitation services.

Neurostimulation techniques are the first treatment option for relief of pain due to peripheral nerve injuries and to limb amputation; if these methods fail, DREZ lesion surgery can be considered for both. In peripheral nerve injuries, DREZ surgery may be particularly useful when the predominant component of pain is paroxysmal (electrical shooting pain), associated with allodynia, hyperalgesia, or both.²⁹ After limb amputation two types of pain may develop and may coexist: pain in the phantom limb and pain in the stump. The former may be reduced by DREZotomy, while the latter is inconstantly influenced.²⁹

Finally, MDT can be performed also for severe occipital neuralgia, unbearable laterocervical pain, and postherpetic pain, but the procedure has been performed only in small subgroups of patients and has had a high recurrence rate.^56,57 Based on the fact that muscle tone was decreased in the operated areas after MDT,^36,58 Sindou applied it also for treatment of harmful spasticity with reasonably good results.^59,60–62

In 1978 Nashold described nucleus caudalis tractotomy/nucleotomy, also known as caudalis DREZotomy, in which DREZ lesion surgery is extended to the trigeminothalamic system for treatment of facial deafferentation pain such as trigeminal postherpetic neuralgia and facial anesthesia dolorosa.^10,63 The results are not as good as those for the other major indications of DREZotomy, but almost 50% of patients may obtain pain relief, and it should be considered that these patients are typically not responding to standard narcotic therapies. For patients without prominent allodynia or hyperpathia, on the other hand, neurostimulation procedures are preferred.⁶

Provided that rigorous selection criteria are applied, DREZ surgery can achieve very good pain control in some intractable painful syndromes. Sindou proposed a schematic decision-making process for neuropathic pain.⁵⁶ Accurate DREZ surgery requires good knowledge of the radicular innervation and of the surgical anatomy of the spinal cord and roots.⁵⁶

Sympathectomy

Sympathectomy means the interruption of the paravertebral sympathetic ganglion chain at a chosen level. Used since the end of the nineteenth century for several diseases,⁶⁴ it is still indicated for the treatment of a wide group of conditions sympathetically maintained and in some cases of vascular diseases.^65–67

The term causalgia (“burning pain”) was first introduced by Mitchell in 1864 to describe a syndrome characterized by burning pain following a partial peripheral nerve injury caused by a high-velocity missile. Its classic triad includes burning pain (associated with allodynia and hyperpathia), autonomic dysfunction, and trophic changes.⁶⁸ The severe form is known as major causalgia and implies always a high-velocity missile injury, and minor causalgia is reserved for less severe forms including also some described after nonpenetrating trauma.⁶⁹ The median and sciatic nerves and the brachial plexus are most commonly affected.

Reflex sympathetic dystrophy (RDS)⁷⁰ consists of many pain syndromes that result from a variety of causes that may or may not include direct nerve trauma. The autonomic nervous system was thought to be implicated, although the “reflexive nature” is still to be demonstrated and the distrophic features are not always present. Sudek’s atrophy and Raynaud’s syndrome have been attributed to sympathetically mediated disorders, too.

In order to reduce confusion the term complex regional pain syndrome (CPRS) was coined to describe a symptom complex and not a particular syndrome or a medical entity.⁷¹ The expression sympathetic maintained or mediated pain embraces a spectrum of conditions in which the main symptom, pain, may be associated with vasospasticity and dystrophic features and can be relieved by interrupting the sympathetic outflow to a body region.^66,72–74

Upper extremity and upper thoracic sympathectomy implies dividing the chain below the first thoracic ganglion and resecting the T2 ganglion either with open surgery or endoscopically^75,76 or percutaneously by chemical, electrical, or radiofrequency techniques.^77,78

Lumbar sympathectomy requires the removal of the L2 and L3 sympathetic ganglia (occasionally L1 and T12 to reach the thigh) usually through a retroperitoneal approach.

Open surgical sympathectomy carries high success rates, up to 100%, and low complication rates, between 2.5% and 5%.^6,51,68,79 Pain recurrence is possible, but its rates vary considerably, ranging between 0% and 33% in many series reported.^64,67,68,80 Conversely, it is well recognized that best results are obtained early in the course of the disease before the development of trophic changes,^64,66,73,74 and therefore, rigorous early assessment by a pain specialist is needed.¹⁰

Sympathetically maintained pain is one of the few pain syndromes in which the first treatment option may include an ablative procedure rather than conservative therapy.

Cordotomy Procedures

Percutaneous Cordotomy

Percutaneous AC is superior to conventional open methods. It is performed with the patient under local anesthesia, requires no incision, and allows excellent functional monitoring prior to completion. In addition, with the patient awake in response to neuromodulation, the functional characteristics of the region of the spinal cord to be sectioned can be studied in more detail. It is also more likely to provide pain relief for symptoms above T5 than the open cordotomy. This technique has been pioneered by Mullan and associates,⁹⁹ and later modified by the use of radiofrequency, image guidance, and functional mapping.^100–102

Contrast material should be administered in the subarachnoid space. This could be done preoperatively (20-30 minutes before) by a lumbar puncture (in this case the patient should be maintained in the Trendelenburg position before the procedure) or at the beginning of the procedure by needle insertion at the C1-C2 level. The patient is placed in a supine position in the computed tomography (CT) unit with the head slightly flexed. Under local anesthesia and intravenous sedation if needed, the needle is advanced from the lateral neck into the C1-C2 intradural space. Lateral fluoroscopy allows identification of the appropriate site to start the needle trajectory toward the spinal canal and the final position of the needle (Fig. 49.2A). Once the dura is penetrated, cerebrospinal fluid (CSF) flow will be evident after stylette removal. At this point, an axial scanogram by CT allows the measurement of the spinal cord and is used to better understand intradural needle positioning. The tip diameter of the needle should be guided anterior to the dentate ligament and posterior to the anterior cord (1 mm anterior for lumbar fibers and 2-3 mm anterior for thoracic and cervical fibers). The needle in its ideal position is nearly perpendicular to the spinal cord (Fig. 49.2B). Once the position of the needle is satisfactory, the electrode is passed through the spinal needle into the spinal canal and the spinal cord. Penetration of the pia results in a brief increase in local pain. Correct positioning of the electrode can be performed by CT scan (Fig. 49.2C and D). Now electrical stimulation is performed to adjust the electrode position. Stimulation at low frequency (2-5 Hz) is used to evoke motor responses. Sensory stimulation is done with higher frequency (100 Hz). Placement of the electrode in the STT and SRT area is confirmed by the occurrence of contralateral sensory phenomena, usually a feeling of warmth or cold; with higher stimulus strength, a painful sensation can be evoked.¹⁰³ When adequate electrode positioning has been performed, the initial lesion is made by using an electrode with a 2-mm exposed tip applying 30 to 40 mA for 30 to 60 seconds.¹⁰¹ After this lesion, the patient is examined to evaluate the area of analgesia obtained. Further lesions can be performed in cases of incomplete results in the target area by increasing the time of passing current or advancing the electrode tip in small increments of 0.5, 1 mm. If the procedure is performed satisfactorily, an ipsilateral Horner’s syndrome usually occurs.

FIGURE 49.2 A, Schematic representation of cordotomy procedure. Insertion of the needle into the C1-C2 level in the supine position (B), position of the needle at the right C1-C2 space in the axial computed tomography (CT) image (C), and final position of the electrode in the axial CT scan (D). Schematic representation shows the final electrode position in the target at the level of the cervical spinal cord and the close anatomical structures.

(Modified from Burchiel KJ, editor. Surgical Management of Pain. New York: Thieme; 2002, Figs. 59.1A (part A), 59.3B (part B), 59.3C (part C), and 59.1C (part D).)

The results of the cordotomy are variable and mainly related to the investigator who is doing the reporting and the criteria used for determining success.

The more recent literature is almost exclusively related to percutaneous procedures. The results are better for pain related to malignancies. In experienced hands, the initial pain control is achieved in 90% of patients. However, the level of analgesia falls with time, reaching 40% of patients at 2 years.¹⁰⁴

The complications resulting from AC are related to the damage of the STT and SRT and of all the other adjacent tracts (Fig. 49.2D). Damage to STT and SRT can lead to painful dysesthesias. Damage to nearby structures can result in (a) respiratory failure and death, (b) bowel, bladder, and sexual disfunction, (c) hypotension, (d) ipsilateral weakness, (e) mirror pain, or (f) ataxia. In cervical cordotomies, respiratory dysfunction can be the principal complication. This is particularly true for patients with preexistent respiratory disorders and in cases of bilateral cordotomies. In this last situation, respiratory functions should therefore be carefully observed in the postoperative period. In the past, with the use of open cordotomy, higher rates of complications were observed. Two reasons can be recognized as the cause of complications: (1) needle mislocalization, which is more difficult with the use of CT-guided AC; and (2) involuntary enlargement of the sectioned area, which is less frequent with the use of recently developed electrodes and small thermic lesions. Therefore, percutaneous AC offers a reduction in the percentage of complications.

Percutaneous CT-guided anterolateral cordotomy is an effective method to control pain in cancer patients. It can be used in association with other less invasive methods, such as morphine pumps and neurostimulation, to offer these patients a higher quality of life.

Cingulotomy

Cingulotomy has been successfully used to treat intractable cancer pain.^105–108 The cingulate cortex is a key component of the limbic system, and its anterior cingulate gyrus plays an important role in the integration of nociceptive, motor, affective, and memory functions.¹⁰⁹ Experimental studies on animals and also functional studies in human volunteers demonstrated a role of cingulate gyrus in control of contralateral nociceptive information.^110–115

Cingulotomy can be considered as an option for patients with intractable cancer pain after failure of antineoplastic and pharmacological treatments and when conservative anesthetic procedures are not available.

Proper pre- and postoperative neuropsychological evaluation should be performed in every case. Magnetic resonance imaging (MRI)-guided stereotactic bilateral cingulotomy is performed.

Results of bilateral cingulotomy are variable. It can assure an improvement of pain control in selected cases.¹¹⁶ However, even after MRI-guided localization of the cingulate cortex, complications related to behavioral changes or cognitive deficit can occur.¹¹⁶

MRI-guided stereotactic cingulotomy, although valuable in selected cases, probably should be reserved for complex and refractory pain problems.

Mesencephalotomy

After the first attempts by Walker¹¹⁷ and Drake and McKenzie,¹¹⁸ the procedure was modified by the application of stereotactic technique¹¹⁹ and avoidance of the medial lemniscus. With these changes, it was possible to reduce the incidence of severe dysesthesia and to reduce the mortality rate.

The first attempts in open surgery were to perform a section of the lateral spinothalamic pathways, but these resulted in unacceptable risks. Stereotactic procedures were therefore applied to interrupt nociceptive pathways with more precision. Looking for an ideal target for pain control, Shieff and Nashold^120,121 elucidated that the procedures worked better and with fewer side effects when the lesion involved the multisynaptic reticulospinal pathways, sparing the spinothalamic tract, which was the original target.

The primary indication for mesencephalotomy is management of cancer pain involving the head, neck, or upper extremities when all other treatments have failed. The most important consideration is patient selection; in fact, this invasive approach can be considered only after an extensive clinical evaluation with a good correlation between the site of cancer involvement and the generation of pain. It can be of particular help when patients report the generation of intense psychological suffering.

Stereotactic procedure is performed to complete the sectioning, following target selection.

Gybels and Sweet¹²² reported pain relief in 86% of patients, usually lasting for the remaining part of the patient’s life. Mortality rates ranged from 1.8% to 8%, although it can be difficult to attribute deaths to the procedure itself and not to the disease. Oculomotor dysfunction is a relatively frequent complication. Dysesthesia can occur in 15% to 21% of cases.

Mesencephalotomy can be a valuable option for the management of cancer pain in selected cases.

Percutaneous Techniques for Trigeminal (and Glossopharyngeal) Neuralgia

The first clear description of trigeminal neuralgia (TN) was given in 1671.¹²³ John Fothergill,¹²⁴ in a paper published in 1773, described the typical features of TN, including its paroxysmal nature and association with triggering factors such as eating, speaking, or touching the face. In 1756, the French surgeon Nicholas André coined the term tic douloureux to describe at least three patients with TN treated by neurectomy.^125–128

The clinical features of the so-called classic idiopathic TN have been well documented and are now universally recognized. One of the pillars in our understanding of this syndrome is the classification that Eller and colleagues proposed in 2005.¹²⁹ These authors classified TN on the basis of two broad categories: the patient’s history and seven specific diagnostic criteria. Following Eller’s classification, TN type 1 identifies the classic idiopathic form, also known as the typical form of TN, and from now on we will refer to TN as this entity. TN is a chronic pain syndrome whose patients suffer from idiopathic episodes of spontaneous facial pain. This pain, which is experienced in one or more divisions of the trigeminal nerve, expresses itself as paroxysms of brief and excruciatingly intense bouts of stabbing, electrical shocks. These paroxysms of TN can arise spontaneously or in response to gentle tactile stimulation of a trigger point on the face or in the oral cavity. They may also be triggered by such natural activities as chewing, speaking, swallowing, touching the face, or brushing the teeth. TN pain is almost always experienced unilaterally, although there have been some relatively rare reports describing bilateral signs and symptoms.¹²⁹ TN type 1 pain typically occurs after conspicuously obvious pain-free intervals that can last for weeks, months, and years. The neurological examination is almost always normal, although cases have been reported in which a slight degree of sensory loss has been described.¹³⁰ The diagnosis of TN is always based on the patient’s clinical history. In terms of its pathophysiology, the features of classic TN are currently thought to be related to a compression of the trigeminal nerve root, usually by a blood vessel, at or near the root entry zone of the trigeminal nerve.^131–134

The incidence of TN is about 2.3 to 4.5 per 100,000 new cases per year.¹³⁵ Age at onset is variable, but the incidence increases in the fifth and sixth decades of life.

From the time of its first recognition, TN has been treated in a variety of ways.^136,137 Today several medical and surgical treatments are available.

Medical treatment is mainly based on the use of an anticonvulsant: carbamazepine, phenytoin, and gabapentin are the most frequently recommended. Centrally acting muscle relaxants, such as baclofen, are also used.

Glossopharyngeal neuralgia (GN) is a spasmodic, lancinating, and paroxysmal pain that starts in the posterior throat or base of the tongue and often radiates down the throat and side of the neck. Often pain is triggered by swallowing or yawning. Typically, cocainization of the area reduces pain and has been proposed to confirm diagnosis.¹³⁸ Seldom, hypotension, syncope, and cardiac irregularities may accompany GN.^139,140 Isolated GN occurs at a rate of 1 case in every 70 to 100 cases of TN;¹⁴¹ concurrence of TN and GN is also possible.¹⁴²

Percutaneous Trigeminal Radiofrequency Thermocoagulation

Even though it had been tried since the beginning of the twentieth century,^151,152 it was not until the mid-1960s, with the modifications of the technique by Sweet, that this procedure became popular. Sweet’s modifications consisted of electrophysiological stimulation for precise localization, reliable radiofrequency current for nerve destruction, intermittent patient sedation with short-acting intravenous drugs allowing intraoperative patient assessment, and temperature monitoring to control precisely the configuration of the site.^94,153,154 With these modifications the use of radiofrequency rhizotomy spread widely and it is still commonly used.

The rationale is based on the fact that pain fibers are carried by unmyelinated C-fibers or thinly myelinated Aδ-fibers that are blocked at a lower temperature than those of larger Aα- and β-fibers carrying deep and tactile sensations.^155–157 Thus, radiofrequency-induced heat will destroy the former fibers, sparing the latter ones.

The procedure is done in the operating room, even though some authors perform it as outpatient surgery in the radiography suite.

The technique consists of the introduction of a needle under fluoroscopy at the level of the gasserian ganglion through the foramen ovale, entering 2.5 cm lateral to the oral commissure and aiming toward the medial aspect of the pupil (Fig. 49.3A and B). Several methods have been described to place the needle into the foramen ovale;^158–160 careful checking under fluoroscopy (Fig. 49.3C) and considerable training and experience are mandatory to avoid internal carotid artery injuries or penetrating the foramina adjacent to the foramen ovale. Once at the target, confirmed by outflow of CSF, the electrode can be inserted.

FIGURE 49.3 A, Schematic representation of the anatomical landmarks for needle insertion in the foramen ovale, following Hartel’s technique. Point 1, medial aspect of the pupil, and point 2, 3 cm anterior to the auditory external meatus, indicate the direction of the foramen; point 3, 3 cm lateral to the oral commissure, defines the point of needle initial penetration. B, Intraoperative picture of the needle insertion toward the patient’s cheek in the right foramen ovale. C, Intraoperative fluoroscopy showing the needle inserted into the left foramen ovale. D, Intraoperative clinical assessment of trigeminal functions after thermocoagulation of left trigeminal nerve.

Electrode placement is guided by fluoroscopy: For V3 pain, the electrode tip lies within 5 mm proximal to the clivus profile and is directed caudal. For V2 pain it should placed at the clivus profile, and for V1 pain the electrode tip lies within 5 mm distal to the clivus profile and is directed cephalad. The electrode tip should not be advanced more than 8 mm deep to the clivus profile because it can injure the abducens nerve in Dorello’s canal or penetrate the temporal lobe. Final placement of the electrode tip is determined by the patient’s response to electric stimulation.

With the patient under mild sedation, an average of three lesions at 60 to 70 degrees for 90 seconds are usually done. The goal is analgesia in the primarily affected trigeminal division, hypalgesia in correspondence of trigger points, and mild hypalgesia in the secondarily affected trigeminal divisions. After each lesion, clinical assessment of facial sensitivity, corneal reflex, and strength of the muscles of mastication is mandatory (Fig. 49.3D).

The procedure does not require general anesthesia, can be performed for patients not suitable for open surgery (the elderly, those with comorbid conditions, etc.), and can be repeated if necessary. Immediate pain relief can be obtained in up to 98% of patients,^147,161 but the long-term recurrence rate may be particularly variable: 25% to 37% at 5 years, 25% to 80% after 10 years,^162,163 or 27% overall (early and late recurrence) in a study with an average follow-up period of 9 years.¹⁶⁴ Almost all patients develop numbness in the face after the procedure, but in most cases this is tolerable. The mortality rate is extremely low;^147,165 loss of corneal reflex occurs in 3.5% to 7%; and cheratitis is seen in less than 3%.^147,148 Dysesthesia develops in 6% to 24% of patients, but anesthesia dolorosa occurs in only 0.2% to 4%, with higher risk in more complete or extensive procedures.^{147,150,161,164} Trigeminal motor function may be impaired, but most of the time this effect is transient.¹⁶⁴ When patients are asked to rate the procedure and the outcome, the majority of them are usually satisfied.^162,164

By the same technique, the glossopharyngeal nerve can be reached percutaneously at the foramen lacerum, allowing a postganglionar sectioning of the fibers.¹⁶⁶

Percutaneous Trigeminal Balloon Compression

The idea of compressing with open surgery the root of the trigeminal nerve to treat TN was first proposed around 1950, but the results were not encouraging.^167,168 Later, Mullan proposed the gasserian ganglion compression through a percutaneus approach, reporting good results.^169–171

The procedure is performed in the same way as for radiofrequency thermocoagulation until the ganglion is reached. Then a balloon catheter is inserted. A characteristic pear figure appears under fluoroscopy when the balloon is properly inflated with contrast agent in Meckel’s cave (Fig. 49.4). The goal is to compress the nerve at 1100 mm Hg or 1.3 to 1.5 atm for 1 minute.

FIGURE 49.4 When the balloon is properly inflated with contrast agent in Meckel’s cave, the typical pear-shaped figure appears under fluoroscopy.

Balloon compression is particularly indicated for first trigeminal division pain, because it carries a lower risk of corneal reflex loss and cheratitis compared to radiofrequency thermocoagulation. Nevertheless, during the procedure the patient may develop bradycardia and hypotension¹⁷² and a pacemaker may be required; the compression is mild and this leads to a higher rate of pain recurrence. The initial rate of pain relief is 92% and the overall recurrence rate is 26%; postoperative numbness and mild to moderate hypesthesia occur in about 50% to 60% of patients; and anesthesia dolorosa and corneal anesthesia are extremely rare, less than 0.1% of cases. On the other hand, masseter weakness can occur in about 10% of patients.¹⁴⁷

Microvascular Decompression for Trigeminal (and Glossopharyngeal) Neuralgia

Dandy¹⁷³ was the fist to propose that there might be a causal relationship between pain paroxysms and the compression of the trigeminal root. This compression is typically caused by adjacent arterial loops, although on occasion, tumors, arteriovenous malformations, and aneurysms located in this region have been known to cause this kind of compression. Though Dandy was the first to describe the role played by compression in 1929, it was not until the 1950s that any kind of therapeutic relief became obtainable by decompressing the trigeminal nerve in TN patients.^174,175 The notion that a microvascular compression might be causing TN has gained much support from the work of Jannetta,^{134,176–178} who was able not only to find a compressing vascular contact in a high percentage of TN patients but also to demonstrate that prolonged pain relief could be obtained by decompressive surgery, without causing any sensory loss. Patients suffering from typical idiopathic TN usually are found to have a blood vessel that is in close contact with the trigeminal root entry zone (TREZ). This finding is typically made during surgery, during the exploration of the cerebellopontine angle (CPA), or radiologically by MRI^{176,178–180} (Fig. 49.5).

FIGURE 49.5 T1-weighted magnetic resonance imaging (MRI) with intravenous paramagnetic contrast agent showing a neurovascular conflict with the left trigeminal nerve (white arrow).

Microvascular decompression (MVD), also known as Jannetta’s procedure, is a nondestructive treatment designed to resolve the compressive conflict between the trigeminal nerve and a blood vessel. Today, it is widely recognized as the primary therapeutic option in the treatment of TN.^{178,181–184}

The procedure is performed in the operating room under general anesthesia. With the patient in the sitting or lateral position, a small retrosigmoid craniectomy is done, allowing for exposure of the transverse sinus, the sigmoid sinus, and the junction between the two. Intradurally, by means of opening the CPA cistern, the CSF drainage facilitates a natural cerebellar relaxation, creating a perfect surgical corridor to the fifth cranial nerve and to the seventh to eighth cranial nerve complex. The TREZ is the most common site of vascular compression. It is the locus of transition from central to peripheral myelin and the most vulnerable part of the nerve. After the neurovascular conflict is identified (Fig. 49.6A), a sharp dissection of the arachnoidal bands that fix the artery into its position allows for moving the nerve away from the artery. Many methods have been advocated for keeping the vessel separated from the nerve, including using pieces of muscle, fascia, periosteum, or subcutaneous fat. Our experience is that these strategies usually have little effect. This kind of tissue is usually reabsorbed and favors adhesions, fibrosis, and arachnoiditis, as has been observed during reoperations in patients in whom neuralgia has recurred.^185–189 Most authors dealing with MVD now use synthetic implants, particularly Ivalon sponge,^{187,190–192} Dacron knitted material,^189,193 or Teflon felt.^{178,193–196} In most of these reports, a small piece of synthetic material has been interposed between the nerve and the vessel. Though they are considered biocompatible, Ivalon, Dacron, and Teflon may, in fact, generate granulomas and carry the consequent risk of distorting or irritating the trigeminal root and, in so doing, producing a recurrence of pain.^197–199 The so-called sling retraction technique^{183,193,200,201} has been proposed as an alternative by introducing the concept of a noncompressive technique for MVD (Fig. 49.6B), with good results.²⁰¹

FIGURE 49.6 A, Intraoperative view of the neurovascular conflict with the left trigeminal nerve, which is stretched by the artery and is almost flattened. B, Intraoperative view showing the resolution of the conflict by noncompressive technique with a small piece of Surgicel (white arrow) that keeps the artery away from the nerve.

The results with this technique are excellent, with up to 98% of patients reporting immediate pain relief and more than 85% and 80% of patients pain-free without medication at the 5 and 9 years’ follow-up, respectively.^150,202 The mortality rate is less than 1%.^150,192,203 The facial numbness rate is only 2%; the rate of dysesthesia, anesthesia dolorosa, corneal anesthesia, and trigeminal motor disfunction approximates 0%. The overall cranial nerve deficit is about 3%;^150,204 fourth and seventh cranial nerve deficit with diplopia and facial weakness may occur, although such effects are usually transient. Deafness has been reported as well. CSF leak, wound infection, and meningitis can occur with an overall rate of 10%,¹⁵⁰ but only a few cases require second surgery or prolonged hospitalization.

Similar results in terms of pain relief and complication rate can be obtained with MVD for GN.^205,206

Last, when negative or equivocal exploration is encountered during MVD, partial trigeminal rhizotomy can be performed. The technique was first proposed by Dandy.²⁰⁷ It consists of a partial section of the nerve in the portion adjacent to the brainstem, starting at its posterior inferior margin. This allows the surgeon to spare most of the facial sensibility, corneal reflex, and motor function. Results are rather good, with more than 80% experiencing pain relief, albeit sensory deficit and dysesthesia are obviously higher than with MVD.²⁰⁸ This procedure can be effective also in patients with demyelinating or neurodegenerative diseases and some consider it the last line of defense for patients who have proved to have a recalcitrant form of TN and have failed all previous treatments.

Deep Brain Stimulation

The first report of deep brain stimulation (DBS) procedures affecting perception of pain dates back to the 1950s, when Pool observed an analgesic effect by stimulating the septum pellucidum and the columns of the fornices during a psychosurgery intervention;²⁰⁹ subsequently, the same author and Heath reported an antinociceptive effect after the stimulation of the paraseptal regions.^210,211 The first report of thalamic DBS for control of pain was by Mazars in 1960.²¹² In the 1970s, Hosobuchi and Richardson were the first to report a case of thalamic and periaqueductal gray (PAG) and periventricular gray (PVG) stimulation.^213–216 According to the gate theory by Melzack and Wall²¹⁷ thalamic stimulation induced paresthesias in the pain-affected areas.

Another cerebral region that led to pain alleviation and also paresthesias when stimulated was the posterior limb of the internal capsule.²¹⁸ The first report of stimulation of the centromedian nucleus of the thalamus is due to Andy.²¹⁹ Although the stimulation of PAG and PVG does not preclude paresthesias, it can lead to “pleasant” sensations in the patient. Another potential target is the centromedianus parafascicular complex of the thalamus, which has been used by Andy for the treatment of dyskinesias.

Many manuscripts discuss the use of DBS in the treatment of drug-refractory pain; nonetheless, an important application of this technique is not possible given the low number of patients considered suitable for it; furthermore, many controversies exist on the optimal target and on the different kinds of pain that could benefit from this therapeutic approach.

The mechanism of action of DBS on pain control depends on the target employed. The thalamus, PAG, and PVG are the most commonly used targets according to the international literature. Some authors believe that the therapeutic effect of PAG/PVG stimulation can be mediated by endogenous opioid substances, on the basis of studies that demonstrate that such an effect can be reversed by administration of naloxone. Other authors instead think that the analgesic effect is due to the activation of several descending suprasegmental systems involved in nociception, included the amine-dependent system.

The analgesic effect of the stimulation of ventralis posterolateralis (VPL) and ventralis posteromedialis (VPM) nuclei, involved in the transmission of sensitive inputs, is, on the contrary, poorly studied and understood. It has been proposed that it can be mediated by the inhibition of neurons of the spinothalamic tract or of the neurons of the dorsal horn. It has also been proposed that a dopamine-mediated mechanism is involved; the most accredited proposal suggests that thalamic stimulation activates the raphe magnus of the rostroventral region of the medulla oblongata, thus leading to the activation of descending inhibitory pathways.

Many clinical observations would tend toward a greater efficacy of the stimulation of PAG/PVG regions in the treatment of nociceptive pain, and of VPL/VPM complexes in the treatment of neuropathic pain; however, as stated earlier, the number of patients treated is still too low, and it is not possible to draw any definitive conclusion without controlled, prospective, and randomized studies. It is also necessary to consider the clinical situations of “mixed” pain (with both nociceptive and neuropathic components) and this once again emphasizes the importance of an accurate patient selection (some authors consider the future use of simultaneous implantation of both thalamic and periventricular electrodes²²⁰).

Painful syndromes that can be treated with DBS include anesthesia dolorosa, postischemic pain, thalamic pain, brachial plexus avulsion, postherpetic neuralgia, postcordotomy dysesthesia, medullary lesions, failed back surgery syndrome, cancer-related pain, and peripheral neuropathic pain.²²¹ Other important indications for DBS (the target being in this case the posterior hypothalamic region, pHyp) are refractory trigeminal autonomic cephalalgias (TACs), including chronic cluster headache (CCH). DBS of the pHyp was the first application in which the choice of target was motivated by neuroimaging functional data. Activation of the pHyp during cluster headache pain attacks was observed during positron emission tomography (PET) studies;²²² this original observation led to the placement of deep brain electrodes within the pHyp to inhibit the pathologically activated neuronal pool in patients with CCH (Fig. 49.7).

FIGURE 49.7 Left: Virtual ventriculographic plates showing the target (pHyp: pink and red) in coronal sagittal and axial sections as part of surgical planning. Different possible trajectories are shown. Right: Postoperative computed tomography (CT) images merged with preoperative magnetic resonance imaging (MRI) showing the definitive location of the electrode in coronal sagittal and axial planes.

Series reported in the literature include 50 patients affected by CCH.^223–229 Five patients affected with TN due to demyelinating disease,²³⁰ two patients affected by short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing (SUNCT),^231,232 one patient affected by chronic paroxysmal hemicrania,²³³ and four patients affected by neuropathic pain of the face²³⁰ have been reported in the literature.

With regard to CCH, in the Belgian group,²²⁸ whose mean follow-up period was 14.5 months, two out of four patients were pain-free, one patient had a dramatic reduction of pain attacks to less than three per month, and one patient had only transient clinical benefit. In the open-phase period of the study reported by the French multicenter study,²²⁶ the mean weekly attack frequency decreased by 48.4%, and 6 out of 11 patients were considered as “responders” (decrease of at least 50% in the frequency of weekly attacks). In the German group,²²³ three out of these six patients were reported to be almost completely attack-free (one mean pain attack per month) after a follow-up period ranging from 9 to 17 months. Of the two patients reported by the British group²³³ one patient only reported infrequent pain attacks (seven injections of sumatriptan) at 11 months’ follow-up, and the other patient reported a decrease in attack frequency from daily to weekly with “massive reduced severity.” In the U.S. group,²²⁹ three out of five patients could be considered responders because of “50% reduction in headache frequency intensity, or both”; the follow-up period ranged from 6 to 12 months in their study. As far as our group is concerned,²²⁷ the mean follow-up time was 4 years; a state of persistent freedom from painful attacks was still present in 10 of 18 patients (62%). Four patients (25%) still required prophylactic drugs to prevent the attacks. In the last 2 years of follow-up, three patients no longer benefited from stimulation despite several changes in parameters. In these three patients, the disease changed from the chronic to the episodic form. Two patients affected by SUNCT who underwent posterior hypothalamic DBS have been reported in the literature to date. The first patient was reported by our group in 2005.²³¹ The preoperative frequency of pain bouts ranged from 70 to 300 per day; after surgery, a complete and definitive remission of symptoms has been confirmed at the last clinical examination, performed at 5-year follow-up. The second patient with drug-refractory SUNCT treated with hypothalamic DBS was reported by Lyons and co-workers in 2009;²³² a 63% reduction in the mean number of daily attacks (133 attacks per day preoperatively vs. 45 attacks per day) was observed in the first postoperative month; at 12-month follow-up a further improvement was observed, with 80% in reduction in frequency of attacks (25 attacks per day).

Only one patient affected by chronic paroxysmal hemicrania (CPH) submitted to posterior hypothalamic DBS has been reported to date.²³³ One of the main criteria for diagnosis of CPH is an absolute response to indomethacin administration; in this patient, however, this drug was discontinued because of the onset of iatrogenic gastritis superimposed on a preexisting Barrett’s esophagus; given the refractoriness of symptoms, the patient underwent pHyp DBS. After a 27-month follow-up period she was reported to be symptom-free. At our institute, we implanted with pHyp DBS three patients affected by secondary neuropathic trigeminal pain (the conditions leading to the painful condition were a posterior mandibular carcinoma with subsequent radical transmandibular tumor resection in one patient, a minor dental procedure in the second, and a nasopharyngeal carcinoma in the third patient).²³⁰ None of the three patients had reduction in painful symptoms at the last clinical examination despite several attempts in modification of stimulation parameters. At our institution five multiple sclerosis (MS) patients affected by refractory TN were submitted to pHyp DBS;³⁹ three patients had beneficial effects within 24 hours from the procedure; all patients reported a reduction of paroxysmal pain attacks within the ophthalmic branch after surgery. Three patients complained of recurrent pain in the second and in the third branch (but, importantly, not in the first) and underwent further radiofrequency thermorhizotomies.

The other two patients reported relief of pain in all three trigeminal branches by combining stimulation with analgesics and did not necessitate further surgical procedures.

The overall percentage of patients with CCH treated with pHyp considered to be responders is 63%.²³⁴

Complications of DBS surgery are mainly hemorrhages, infections, failure of the components of the stimulation systems (including breakage of the leads), and stimulation-related side effects, which depend on the target being stimulated (seizures, diplopia, dizziness, paresthesias, cephalalgia).

Motor Cortex Stimulation

The first report about the stimulation of the motor cortex is due to Penfield, who observed some sensitive responses after the stimulation of the motor cortex in a patient previously submitted to postcentral gyrus removal during a surgical procedure for the treatment of epilepsy. Penfield and Jasper subsequently treated refractory hemisomatic neuropathic pain through contralateral postcentral corticectomy; further removal of adjoining precentral gyrus was used to reclaim clinical benefit when a relapse of symptoms occurred.

White and Sweet and Lende performed pre- and postcentral corticectomies for the treatment of cases of hemisomatic and neuropathic facial pain. These reports constitute the background for subsequent employment of motor cortex stimulation (MCS); this research was motivated by the controversial results of DBS and by the progressive refusal of ablative procedures (corticectomy). Although the stimulation of the sensitive, rather than motor, cortex would be in accordance with the “gate” theory of Melzack and Wall, several reports of stimulation of cortical areas different from the main somatosensory area (SI) (such as frontomesial cortex stimulation, performed by Hardy in 1985, which resulted in elevation of the latency of the nociceptive response) allowed an initial understanding of the complexity of the cortical pain modulation mechanisms. It must be remembered that in 1986 Hosobuchi reported the effectiveness of the stimulation of sensitive cortex for the treatment of pain in the lower limbs.

In 1991 Tsubokawa introduced the epidural motor cortex stimulation for the treatment of central deafferentation pain, and his group concluded that sensitive cortex stimulation was not effective for this purpose. He revealed that such technique was able to inhibit burst-like thalamic discharges that had been previously demonstrated to be altered in animal models of deafferentation pain. He also demonstrated the safety of this procedure with particular regard to eventual postoperative development of seizures.²³⁵

The mechanism of action of MCS is still poorly understood; the comprehension of its physiological effects strictly depends on that of the anatomical connections between the motor cortex and several cortical, subcortical, brainstem, and spinal cord structures. These connections can now be revealed through animal studies and different neuroradiological studies, such as positron emission tomography (PET) and single-photon emission computed tomography (SPECT). Some SPECT studies demonstrated a normalization of the local cortical somatosensory-motor (SI-MI) circuit²³⁶ after MCS; this observation could be in line with Tsubokawa’s original view, according to which MCS activates non-nociceptive neurons of primary sensory cortex through both orthodromic and antidromic axonal pathways interconnecting MI and SI, thus leading to activation of surrounding nociceptive inhibition in SI.

Other studies demonstrated that cortical stimulation is able to increase the blood flow to the ipsilateral thalamus, thus modulating the long reverberating thalamocortical loop, and overriding the pathological phenomena of disrupted oscillation and temporal synchronization.²³⁷

Interaction of motor cortex with sensory thalamus, perigenual cortex, dorsolateral prefrontal, orbitofrontal cortex, periaqueductal gray matter (PAG) also seems of relevance; it has been suggested that MCS acts through descending inhibitory controls involving the connection between the anterior cingulated cortex (ACC) and PAG; this could in part explain the positive effect of this treatment modality on both the affective and the evoked components of pain (allodynia and hyperalgesia).

Several painful conditions have been treated with MCS up to now, such as thalamic pain (postictal or posthemorrhagic), postischemic bulbar pain which accompanies Wallenberg’s syndrome, neuropathic facial (trigeminal) pain, phantom limb pain, deafferentation pain, postherpetic neuralgia, spinal cord injuries, and pain due to avulsion of the brachial plexus. One of the first reports of the efficacy of MCS for treatment of central pain is by Tsubokawa (“good” control of symptoms with MCS in 65% of patients with thalamic and putaminal postischemic and posthemorrhagic pain) and by Katayama who performed MCS in three patients who had been submitted to DBS of the VPL nucleus of the thalamus for pain related to Wallenberg’s syndrome; after DBS, these patients reported a worsening of symptoms, but they presented a 60% improvement after MCS. Treatment of facial neuropathic pain has been indicated as the major indication for MCS, in part because of the relatively large cortical representation of the face; and in fact, different authors report a 60% pain improvement after a mean follow-up period of 1 year.^238–242

MCS can be performed under general or local anesthesia, and implies the positioning of plate-shaped electrodes above the dura mater overlying the motor cortex, although some authors place them subdurally; in some centers the electrode is positioned parallel to the motor strip, and in others it is positioned perpendicular to it, with the most anterior contacts located above the precentral gyrus.

The procedure is carried on after the localization of this structure with neuroradiological methods, intraoperative somatosensory evoked potentials (SSEP), intraoperative stimulation, and implementation of functional MRI in the neuronavigation system, expecially in patients with cerebral ischemia, which could lead to plastic rearrangement of the cortex itself.

Some authors perform MCS by sliding the plate electrode through a simple bur hole centered on the posterior portion of the previously localized precentral gyrus, whereas others perform a craniotomy. Once positioned, the abovementioned integrative methodologies can be used, thus permitting a refinement of the positioning according to the case. In general, it is preferable to position the active contacts just above the area of the motor strip representing the affected part of the body (Fig. 49.8) given that the lower limb representation lies in the medial surface of the hemisphere (anterior portion of the paracentral lobule). MCS is usually performed for facial or upper limb pain; nonetheless, reports exist of placement of the electrode subdurally over the paracentral lobule, or epidurally, very close to the sagittal sinus; in these cases, 6 out of 12 patients with lower limb pain improved by 40% to 50%.²⁴³

FIGURE 49.8 Upper panel: Three-dimensional (3D) postoperative reconstruction showing the placement of the chronic leads over the motor cortex. Lower panel: Intraoperative stimulation trials evoked electromyography (EMG) response in the abductor pollicis brevis muscle but not in the remaining muscles (from top to bottom: orbicularis oris muscle; brachioradialis m., abductor pollicis brevis m., rectus femori m., anterior tibialis m.).

After the procedure, the implanted pulse generators connected to the intracranial electrodes are activated at various settings, which vary according to the institution, the physician’s preferences, and the results obtained. They range from 1 to 5 V, 15 to 130 Hz, 60 to 450 microseconds. The configuration more often used is bipolar rather than monopolar; when in bipolar mode and in those cases of electrodes placed perpendicular to the central sulcus, the most anterior contacts are set as cathodes.²⁴⁴

It is difficult to systematically assess the overall results of MCS for chronic pain reported in the literature because of differences among indications, surgical procedures, evaluation scales used, and stimulation parameters. Moreover, only a few studies are prospective or crossover placebo-controlled, even though MCS usually does not induce any perceptible sensations and could thus be ideal for such study design. Some authors consider responders to be patients who report at least 50% pain relief, and others consider a 40% decrease in intensity of pain a responsiveness criterion.

The pain evaluation scales most often employed are the visual analog scale (VAS) and the McGill pain questionnaire, although the majority of authors assess the percentage of pain relief or use indicative terms such as “poor,” “fair,” “good,” or “excellent” when referring to the clinical outcome.

As reported by Fontaine and associates²⁴³ only three studies, comprising 10 patients, reported postoperative CT or MRI for determining the exact location of the electrode;^239,244,245 in seven of these patients, postoperative imaging was obtained to rule out electrode misplacement and in the other three the electrodes were thought to be located in the preoperatively determined target with neuronavigation systems and then assessed with intraoperative stimulation.

Among the abovementioned indications, the best results have been obtained in trigeminal neuropathic pain, phantom limb pain, and spinal cord injury; in this group of patients, at 1-year follow-up 50% were considered responders. Despite the clinical, procedural, and assessment incongruity among the various studies, it could be said that approximately 57% of the treated patients benefited from the procedure (40-50% of pain relief); the most frequent adverse effects were related to the system implanted (infections, hardware failure); less frequently, local pain, subdural or epidural hemorrhage, and intraoperative seizures induced by stimulation were observed. In no case were long-term seizures reported.

Spinal Cord Stimulation

Spinal cord stimulation (SCS) was first introduced (with the name stimulation of the dorsal column of the spine) by Shealy in 1967;²⁴⁶ initially, though, satisfactory and definitive results were not always achieved, partly because of continuous and indiscriminate use of the method and also because of poor choice of indications.

Better results were obtained beginning in the 1970s, when more accurate patient selection and advancement in technological equipment allowed this technique to gain a wider acceptance.²⁴⁷

The procedure consists of the positioning of an electrode (with various shapes, dimensions, and number of contacts) in the epidural spinal space; the site of the implant depends on the site of pain. It is generally placed at the cervical or middle-lower thoracic level (Fig. 49.9). The internal pulse generator (IPG) is positioned in the subcutaneous tissue (generally at the abdominal level) and connected to the electrode through tunneled extension leads.

FIGURE 49.9 Intraoperative picture showing the spinal epidural insertion of the electrode for spinal cord stimulation; in the insert, different configurations of leads available for such procedures.

Stimulation is actually carried out at the level of the posterior funiculi of the spinal cord, where all large-caliber sensitive fibers are grouped. The electrode can be implanted with a laminectomy or with percutaneous technique; in some centers the definitive implant of the system is preceded by a trial stimulation period, during which the clinical efficacy of the device is evaluated with an external stimulator. It is now possible to have programmable devices, which are given to the patient, who can use them for starting or stopping the stimulation with an external magnet that communicates with the IPG. Stimulation parameters employed range from 1 to 5 V as amplitude, 30 to 80 Hz as frequency, and 100 to 300 microseconds as pulse width; in electrodes with complex arrangements of the contacts, the configuration of the electrical field plays a major role in clinical outcome.

SCS is actually employed mostly for neuropathic rather than for nociceptive pain; such pain control does not affect other sensory modalities. Generally after the implantation the patient should perceive paresthesias in the same region of the pain; these paresthesias are due to the stimulation of large myelinated fibers in the dorsal column.

Indications for SCS are failed back surgery syndrome, complex regional pain syndrome (I and II), incomplete avulsion of brachial or lumbar plexus, phantom limb pain, postherpetic neuralgia, pain secondary to spinal cord trauma, peripheral vascular disease in the lower limbs, and angina pectoris. Obviously, all these conditions must be definitively drug refractory to take into consideration SCS as a therapeutic option.

The mechanism of action of SCS is still debated; numerous theories have been proposed, the first of which has obviously been the “gate” theory of Melzack and Wall. New physiological mechanisms have been hypothesized, such as the “orthodromic theory” (the electrical impulses conveyed by the fasciculum gracilis and cuneatus activate suprasegmental centers such as PAG and nucleus of raphe magnum, which exert inhibitory control on pain perception via descending pathways), the “collision current theory” (the current traveling along the dorsal roots enters in collision with orthodromic stimuli traveling through Aδ- and C-fibers, blocking them), and the “inhibitory neurotransmitter” theory, according to which the analgesic effect could be due to the stimulation-induced release of endogenous opioids at the level of the dorsal horns, or to the fact that stimulation leads to an increase of the concentrations of neuroactive substances in CSF such as norepinephrine, GABA, substance P, adenosine, and glycine. In particular, the release of inhibitory amino acids such as GABA and glycine could limit the release of excitatory amino acids such as aspartate and glutamate and could act on their own receptors (especially GABA_B receptors) at various spinal levels, restoring normal GABAergic activity within the spinal cord.

Probably different mechanisms play a role in the analgesic effect of SCS, and as stated by Linderoth²⁴⁸ the final common effect could be a suppressive effect of current on wide-dynamic range and hyperexcitable neurons located on the dorsal horn of the spinal cord.

As far as ischemic pain is concerned, the analgesic effect of SCS could instead be due to a rebalance between request and contribution of oxygen, increasing oxygen contribution and decreasing its requirement. Furthermore, SCS could lead to peripheral vasodilatation by antidromically activating Aδ fibers traveling in the dorsal roots, with subsequent release of vasoactive substances such as calcitonin-gene-related peptide (CGRP); also, it could inhibit the sympathetic efferent activity, reducing its vasoconstrictive effect both at the level of the lower limbs and at the cardiac level.

Most of the studies report a clinical efficacy of SCS in 50% to 70% of the patients treated (responders: reduction of at least 50% in intensity of pain). SCS has been reported to be superior to surgical reexploration in failed back surgery syndrome in a study by North and co-workers.²⁴⁹

The most frequent adverse effects reported in the literature are lead displacement, fractures of the extension leads, infections, CSF leaks, and epidural hematoma. The rate of displacement of cervical leads is largest in cervical electrodes because of the greater motility of the cervical spine.

Peripheral Nerve Stimulation

Whereas large myelinated fibers belonging to nervous structures located upstream from the site of the pain are the target of SCS, regional subcutaneous tissue, where small-caliber fibers run (also called “unnamed fibers”), or single nervous trunks, are the targets of peripheral nerve stimulation (PNS).

Nowadays there is some confusion about this terminology, as in the case of occipital nerve stimulation (ONS), in which the electrode is not just positioned above the nerve but in the proximity of its anatomical course below the inion; this kind of stimulation could be considered an “indirect PNS” rather than subcutaneous stimulation.

The first report of PNS for pain dates back to 1967, when Wall and Sweet inserted an electrode into their own infraorbital foramina and observed a decrease in painful sensitivity during the period of stimulation.²⁵⁰ In the same period Shelden implanted electrodes around the mandibular nerve in three patients and delivered electric current through an implanted receiver at a frequency of 14,000 Hz; again, temporary pain decrease was obtained.²⁵¹ Weiner and Reed in 1999 described the first percutaneous insertion of an electrode in proximity to the occipital nerves.²⁵² This report paved the way for the clinical application of ONS, which is now employed for several refractory painful conditions. Slavin and Burchiel subsequently performed the PNS approach for refractory pain in the trigeminal and occipital territories.²⁵³ Subsequently, other reports were published reporting the efficacy and different surgical techniques of this procedure in different syndromes.^254–256

In general, the relevant teleological difference between the procedures aimed at interacting with the central nervous system interface and PNS consists of the stimulation of areas that are known to be unsatisfactorily controlled by SCS or by major trunk stimulation. In fact, with PNS the electrodes are typically placed above or very close to the area of perceived pain. This procedure is usually performed with cylindrical leads rather than paddle-type leads, in order to resemble a circumferential electrical field.

The basic principle of PNS is the vicinity of the placed electrodes to the area of referred pain. The characteristics of pain should dictate the appropriate placement of the electrode(s); as stated by Barolat,²⁵⁷ in case of mild allodynia, the implanted electrode or electrodes should be still better placed in the painful area; in cases of severe allodynia, the electrodes should be placed on each side of the allodynic region, so as to “circumvent” it. Barolat states that with this technique allodynic symptoms are likely to be significantly reduced.

The mainstay of the subcutaneous electric stimulation is, as stated previously, to relieve pain in areas that are difficult to reach by SCS or DBS, such as median lumbar pain, median thoracic pain, inguinal pain, or scapular pain, by directly reaching and intercepting the involved peripheral nervous components that generate the aberrant nociceptive discharges. The PNS procedure can be performed under general or local anesthesia, according to the extension of the painful area and compliance of the patient to be treated, and the definitive implant can be preceded by a trial stimulation period by means of external pulse generators delivering appropriate current intensity at established stimulation parameters; in case of satisfactory pain relief in the trial period, the whole system is definitively implanted.

The most frequent indications for PNS are axial lumbar pain (which is typically not well controlled by SCS), thoracic pain, shoulder pain, atypical facial pain, inguinal pain, and more extensively, several headache conditions. According to Barolat²⁵⁷ the preferred location of the implanted electrodes should be the superficial layer of the subcutaneous tissue, because a more superficial location could lead to lack of perceived paresthesias (considered to be essential, at least as predictive factors, for effective pain relief, also taking into account the “gate control” theory) and a deeper location could lead to undesired muscle contractions (in cases of proximity of the lead to the muscular fascia). The usual preoperative and noninvasive techniques employed before the operation are TENS and anesthetic peripheral nerve block; nonetheless, there is no consensus about the predictive values of such modalities with regard to subsequent PNS.

PNS is a relatively simple technique performed by tunneling a cylindrical electrode through the skin to the affected area of the body and connecting it to an internal pulse generator. The difficulty of the procedure depends on the amount of adipose tissue of the subject and on the extension of the painful area; furthermore, the complexity and the different dysesthetic features of given regions in a given patient can require careful preoperative planning, taking into account bony prominences and possible sites of skin erosion or lead decubitus; different trajectories can in fact be required to address these important operative issues.

Occipital Nerve Stimulation, Peripheral Trigeminal Stimulation, and Vagal Nerve Stimulation

ONS consists of stimulating the great occipital nerve (GON) by means of implanted paddle-type or wire-type electrodes that are positioned in the suboccipital region under the subcutaneous tissue and above the splenium fascial plane. One or two electrodes can be positioned according to the symptoms and to the patient’s and clinician’s preference. The configuration of such electrodes varies according to the different numbers of contacts and different distance between them. Bilateral implanting procedure implies the use of two symmetrical electrodes positioned about 1 cm below the inion and extending from 1 to 5 cm laterally to the midline, to ensure proper coverage of GONs, which emerge from the splenium muscle’s fascia at this level (Fig. 49.10). As for other PNS procedures, the implantation is completed after positioning of the IPGs in the subcutaneous tissue, more often in the infraclavicular or paraumbilical regions. In several centers a trial-stimulation period is performed before definitive implant of the system; during this period, the electrodes, which are sutured to the skin in a sterile fashion, are connected to external pulse generators. Patients reporting more than 50% decrease in pain intensity or painful attack intensity are considered “responders” and recommended for implantation of IPGs.

FIGURE 49.10 Left: Drawing showing the ideal placement of the occipital lead, from 1 cm to about 5 cm from the inion (midline). Right: Postoperative computed tomography (CT) scan of a patient submitted to occipital nerve stimulation (ONS).

The main indications for performing ONS are migraine,^258,259 transformed migraine,²⁶⁰ cluster headache,^261,262 hemicrania continua,²⁶³ cervicogenic headaches,²⁵⁵ and occipital neuralgia,²⁶⁴ the latter being the most common so far. Slavin and colleagues²⁶⁴ reported on 14 consecutive patients with intractable occipital neuralgia treated with ONS. Ten patients were considered responders after the trial stimulation period and proceeded with system internalization. At the last clinical follow-up examinations, seven patients (70%) continued to benefit from the procedure and were able to reduce the daily intake of prophylactic drugs, whereas three patients had their hardware removed because of loss of beneficial effect or occurrence of infection.

Matharu²⁵⁸ reported results in eight patients submitted to ONS for chronic migraine, who all described the clinical outcome at least as “good,” with pain reduction ranging from 75% to 90%.

Popeney and Alo²⁶⁰ reported on 25 patients with transformed migraine implanted with C1 through C3 peripheral nerve stimulation. The evaluation was carried out using the migraine disability assessment (MIDAS); symptoms had been refractory to conventional treatments for at least 6 months in this series. Prior to stimulation, all patients experienced severe disability (scored as grade IV on the MIDAS). After stimulation, the average improvement in the MIDAS score was 88.7%, and all of the patients reported that their headaches were well controlled after the implantation procedure.

Rodrigo-Royo and associates²⁵⁵ reported on the results concerning four patients with drug-refractory, long-persistent, and severe pain in the occipital region, who were treated with electrical stimulation of the C1, C2, and C3 peripheral nerves. Results were considered good in all patients; the authors observed the disappearance of continuous pain in all patients and the improvement in intensity and frequency of pain bouts, with subsequent reduction or discontinuation of pharmacological therapy. Burns and colleagues²⁶³ reported on six patients suffering from hemicrania continua submitted for ONS. At long-term follow-up, four of the patients reported 80% to 95% improvement, one patient reported 30% improvement, and one patient reported worsening of symptoms.

Bartsch and Goadsby²⁶⁵ in 2003 reported that stimulation of nociceptive afferent C-fibers of the dura mater of the rat can lead to a sensitization of second-order cervical spinal neurons. The authors proposed that such a mechanism could be involved in the referral of pain from trigeminal to cervical structures in patients affected by CCH and chronic migraine. Piovesan in 2003²⁶⁶ highlighted the concept of “trigeminocervical complex,” thus hypothesizing the convergence of nociceptive information in trigeminal and cervical territories at the level of the neurons in the trigeminal nucleus caudalis, extending to the C2 spinal segment. Based on this study, Magis and co-workers²⁶¹ reported on the results of eight patients with refractory CCH submitted to ONS. Two patients were pain-free after a follow-up period of 16 and 22 months; three patients reported 90% reduction in attack frequency, and two patients reported improvement of about 40%. Switching off the IPGs led to recurrence and increase of attack frequency in all the patients. The mean time elapse from implantation to the onset of beneficial effects was reported to be 2 months.

Burns²⁶³ reported 14 patients with drug-refractory CCH submitted to bilateral implantation of ONS electrodes; the median follow-up period was 17.5 months, and 10 out of 14 patients benefited from the procedure. Three patients reported improvement of 90% or better, three reported a moderate improvement (40% or better), and four reported 20% to 30% improvement. Most of the patients reported relapse of pain attacks when the device was switched off; interestingly in one patient the ONS system helped in aborting acute attacks.

Slavin²⁶⁷ reported on a case series of 30 patients submitted to a trial period of peripheral nerve stimulation for refractory craniofacial pain, of whom only 22 experienced more than 50% improvement in pain intensity and were subsequently implanted with the complete system. Three patients received infraorbital subcutaneous electrodes, four received supraorbital subcutaneous electrodes, 13 received occipital electrodes, one patient had a combination of infraorbital and occipital stimulation system, and one patient had a combination of supraorbital and occipital stimulation. Follow-up evaluations were carried on for a mean time period of 35 months. At the last clinical examination, the systems had been removed in five patients because of improvement in pain intensity, persistence of symptoms, and infection; 17 patients continued to use the PNS device; three of them had an improvement in pain of less than 50%, whereas 14 reported an improvement of 50% and were considered responders. Reed²⁶⁸ reported seven patients affected by refractory chronic migraine headaches who received a combined occipital nerve–supraorbital nerve neurostimulation system. In this study, the relative responses to the ONS system alone and to both ONS and supraorbital systems were evaluated. The follow-up period ranged from 1 to 35 months; all patients reported full therapeutic response only to combined supraorbital and occipital neurostimulation, whereas only a partial response was provided by the ONS system alone. The authors conclude that in patients affected by chronic migraine the clinical benefit may be better in those implanted with both systems.

Migraine is often comorbid with a large spectrum of diseases, including epilepsy, stroke, and above all, psychiatric disorders,²⁶⁹ especially major depression; depressive episodes can influence pain’s processing and perception and are associated with a poorer clinical outcome in cephalalgic patients. Given that vagal nerve stimulation (VNS), approved by the U.S. Food and Drug Administration as an adjunctive treatment for drug-refractory depressive disorder,²⁷⁰ has been shown to be effective in controlling episodic migraine,^271,272 our group²⁷³ published preliminary data about the employment of VNS for refractory chronic daily headache (CDH) with comorbid depression.

The vagal nerve stimulation is carried on by wrapping a spiral electrode around the cervical portion of the left vagus nerve. The left nerve is chosen because of the lesser influence on cardiac activity. The IPG is positioned in the left subclavicular region.

In our case four patients were selected for implantation of this device; all of them had a clinical history of CDH lasting at least 2 years, were refractory to conservative treatments for both headache and major depression, and suffered from severe disability on activities of daily living. None of the patients had psychotic features, heart disease, or lung diseases that could contraindicate chronic stimulation of the left vagus nerve. The intensity of stimulation used ranged from 1 to 2.25 mA. Two of the patients showed a significant improvement for both headache and depression after 1 to 3 months from the start of stimulation. The remaining two patients showed limited or no clinical benefit after VNS. Another study by Mauskop²⁷⁴ reported improvement after VNS in two patients with CCH and in two patients with chronic migraine.

Systems of Spinal Intrathecal Release of Opioids

Intrathecal (IT) delivery of analgesic agents allows for a more focused action of the drugs at the spinal cord level (where respective receptors are present at different locations), thus limiting the adverse side effects of systemic administration. Moreover, in this procedure the dosage needed for the analgesic effect is lower than that of intravenous, intramuscular, or oral administration, and this lower dose also limits the side effects.

The discovery of opioid receptors at the spinal cord level paved the way for this procedure; in 1976 Yaksh and Rudy obtained a reduction of the hind-limb withdrawal reflex in rats exposed to algic stimuli and concomitantly submitted to subarachnoid administration of opioids.²⁷⁵ The analgesic effect of opioids at the spinal level is thought to be mediated by inhibitory interneurons that act to reduce the nociceptive transmission at higher levels.²⁷⁶ Intrathecal opioid (morphine chlorhydrate or hydromorphone) therapy is usually performed for pain of nociceptive type, whereas neuropathic pain is sometimes considered a secondary indication, both because of the efficacy of SCS and for the higher opioid concentrations required in this case. Nonetheless, there are reports about the possible efficacy of IT opioids in neuropathic pain.²⁷⁷ Other types of agents are available for this indication (even if still less studied and tested than morphine).

Classical indications for opioid IT therapy are failed back surgery syndrome, cancer pain, axial spinal pain, complex regional pain syndrome, arachnoiditis, poststroke pain, and spinal cord injury. Other prerequisites are the compliance of the patient and the possibility of performing regular clinical follow-up examinations, which are mandatory for the variation of the daily dosage.

Nonopioid drugs, as mentioned earlier, have been proposed, both alone or in combination with morphine, for IT therapy. They include clonidine, tizanidine, ziconotide, glutamate antagonists, benzodiazepines, tricyclic antidepressants, butamin, and nitric oxide synthetase inhibitors.²²¹ Most of these drugs have been tested only on experimental models. Hydromorphone can be used in cases of tolerance to morphine, and IT clonidine can be added to IT opioids in cases of tolerance development or if unsatisfactory results are obtained with one drug alone; in particular, two early studies report the safety and efficacy of this combination in noncancer and in cancer pain.^278,279 Baclofen and bupivacaine can be used as a third-line approach, too, alone or in combination to first-line agents.²⁸⁰

The implantation procedure can be preceded by a trial bolus of IT morphine (generally 0.5 mg is administered at this stage), after which pain decrease is assessed; if no adverse effects appear in case of absent benefit, the bolus dosage can be slightly increased and the patient evaluated again.

Surgery is performed by first positioning a Tuohy needle in the spinal lumbar subarachnoid space, and through it a silicone catheter is introduced; this catheter is then tunneled subcutaneously and connected to a drug’s controlled-release device (pump). The pump is also positioned subcutaneously, and can be externally programmed with a data transferring and receiving device, thus regulating the daily dosage of the administered drug. Different kinds of pumps exist, including pulsatile pumps (hand-operated systems that are activated with tactile stimuli), constant flow pumps (in which case the effective daily dosage of the drug is determined by the concentration of the solution present in the reservoir), and programmable pumps. Programmable pumps are the most versatile devices in that they can be remotely controlled by an external programmer, dictating the daily dosage of the drug according to the concentration of the solution initially present and to the flow rate through the catheter.

For morphine chlorhydrate, initial daily dosage is usually 2 to 3 mg, but it can be varied according to the patient’s needs and to the eventual onset of side effects (dosage spectrum is quite wide, ranging from 1 to about 60 mg/day). For clonidine, the recommended dosage range is 10 to 10,000 µg per day.²⁸¹ In the current literature, the percentage of pain relief is reported to be 60% to 80% for malignant pain and 60% to 65% for nonmalignant pain.²⁸¹

The most common complications of surgery are hardware failure, infection of the surgical wounds, epidural abscess, and formation of seroma at the tip of the catheter. Drug-related complications with opioids are constipation, urinary dysfunction, nausea, hypogonadism, sweating, weight gain, and respiratory depression. The incidence of respiratory depression (requiring anesthesiological intervention) is estimated to be around 1%. The most common clonidine-related adverse effects are hypotension, decrease in heart rate, and mild sedation.

Systems of Intraventricular Release of Opioids

In local or general anesthesia, after precoronal craniectomy, a right intraventricular catheter is positioned (the exact position may vary) and subsequently connected to a subcutaneous access site (reservoir). At this site, injections of morphine chlorhydrate can be delivered, thus obtaining a powerful and long-lasting antalgic effect, through direct action on the morphinic periventricular receptors. This technique is generally used in cases of diffuse cervical and facial pain, often secondary to neoplastic ear, nose, and throat or stomatological lesions and not well controlled by adequate dosages of oral morphine. Also patients affected from Pancoast-Cuffini syndrome can benefit from the procedure. The initial dosage is inferior to that administered intraspinally (0.5 mg in 24 hours) but in this case it can be increased according to the patient’s needs. Despite the central administration, no case of respiratory depression has been reported, thus allowing the drug to act only on the painful symptoms in a selective fashion.