Ablative Neurosurgical Techniques in the Treatment of Chronic Pain: Overview Timothy R

Home , Rhizotomy

Chapter 50 Ablative Neurosurgical Techniques in the Treatment of Chronic Pain: Overview Timothy R. Steel and Kim J. Burchiel

Conservative nonsurgical treatment, including rehabilita- clude neuroma pain, postrhizotomy pain, anesthesia do- tion therapy, is the mainstay of the treatment of most pa- lorosa, postherpetic neuralgia, sympathetically maintained tients with chronic pain, whereas neuroablative surgery is pain, complex regional pain syndrome (CRPS) I and II, pain indicated only in a minority. It is the selection of the patients following brachial plexus avulsion, spinal cord injury pain, who will benefit from ablative procedures that is critical. and the thalamic pain syndrome. The pain often is described Appropriate management of patients with chronic pain re- as burning or crushing. Shocklike exacerbations occur and quires a combined approach between pain specialists, reha- may be spontaneous in most deafferentation syndromes but bilitation teams, and, when necessary, neurosurgeons. Some also may be triggered by cutaneous stimulation, as seen with patients will not be candidates for surgery at any time, causalgia or sympathetically maintained pains. Neuropathic whereas others may become candidates after maximal con- pain is incapacitating and is usually refractory to all conserv- servative therapies have been exhausted. ative measures. Narcotic analgesics characteristically provide The exact time course for pain to be classified as chronic is only minimal relief. Anesthetic blockade proximal to the le- a matter more of convention than of science. Chronic pain sion may relieve deafferentation pain temporarily; however, has been described as that which persists beyond 1 month neuroablative procedures rarely provide lasting relief and after the normal course of an acute illness or after a reason- may cause an even more widespread deafferentation pain able time for an injury to have healed.1 The International As- state. In selected conditions, good results can be achieved by sociation for the Study of Pain stated that chronic pain is destructive procedures. The poor results of sensory pathway any pain that persists beyond the “normal time of healing.”2 ablation suggest that structures other than the spinothalamic Bonica suggested that 3 months is the best time of separa- tract and somatosensory relay nuclei of the thalamus are in- tion between acute and chronic pain.1 Because long-standing volved in the pathogenesis of deafferentation pain. pain experienced by patients with cancer or other chronic Although surgical ablative procedures can be effective in disease states becomes an illness in itself, the pain syndrome the management of chronic pain, they often are discarded as warrants treatment as a specific disorder on its own, no mat- a treatment option by many physicians because historically ter what the time constraints. these procedures are associated with a relatively high com- Chronic pain can be divided broadly into two kinds: no- plication rate. These complications include the development ciceptive pain and neuropathic (deafferentation) pain. Noci- of new deficits as a result of the ablative procedure and mor- ceptive pain results from stimuli that damage or severely bidity associated with operative procedures on high-risk pa- deform tissue and cause the conscious experience of pain. tients, such as those with end-stage cancer. In patients un- Trauma, inflammation, visceral distention, and neoplastic dergoing ablative surgery for benign pain syndromes, the infiltration all cause nociceptive pain. Narcotic analgesics long-term difficulties with pain recurrence and the develop- generally provide relief for such patients, although the po- ment of deafferentation pain also have contributed to thera- tential for these patients to experience tolerance and habitu- peutic nihilism in this area. ation is high. Anesthetic blockade, somatic nerve or root sec- Over the last two decades, the administration of in- tion, and spinothalamic tract section typically will abolish traspinal and intraventricular morphine has both improved nociceptive pain. and simplified the management of pain from many malig- The distinguishing features of neuropathic pain are shared nant processes.3 Direct delivery of small doses of opiates to by many different chronic pain syndromes associated with spinal and ventricular cerebrospinal fluid (CSF) provides ef- peripheral or central nervous system (CNS) injury. These in- fective analgesia in most patients without the major side 633

Surgical Management of Pain. Burchiel. Thieme Medical Publishers, Inc., New York © 2002 634 SURGICAL PROCEDURES effects of sedation, nausea, or constipation. Reported com- tients who survive for longer periods (>6 months), tolerance plications are rare, with more than 90% of patients reporting may develop, requiring greater doses, which may cloud con- significant benefits until death or last follow-up.3 Drug de- sciousness and impair judgment.3 It is these patients who will livery systems of varied complexity and expense are avail- benefit most from neuroablative procedures.7 able. Few patients are truly too ill to undergo the minimally If the neurosurgeon can offer selective neurodestructive invasive surgery necessary for implantation. procedures with minimal or at least acceptable morbidity, This chapter first reviews the general principles of pa- and ideally no mortality, neurodestructive procedures can tient selection for neuroablative procedures and then fo- reliably improve the quality of life for these patients. In cuses on the procedures currently available. We shall dis- some cases, this means the patient can be narcotic free cuss which procedures are appropriate to consider in the for substantial periods of his or her remaining life. In cases management of specific pain syndromes and where in the of narcotic-resistant pain, limited destructive neurosurgical treatment paradigm they should be applied. For most con- procedures often can provide pain relief for the duration of ditions, ablative techniques still are considered second-line the patient’s life.3 The estimated length of survival of these therapy. Only in a minority of situations would an ablative patients is important in determining a treatment strategy. surgical procedure be considered the primary treatment. Loeser believes that patients with less than 3 months’ life expectancy should not be considered surgical candidates because of their higher risk of significant morbidity and GENERAL CONSIDERATIONS mortality.7 In general, patients with expected survival of more than 6 months constitute better candidates for surgical To consider a patient for an ablative procedure, conservative procedures. They typically have lower surgical risk, and the nonsurgical therapy must have been attempted, maximized, analgesia that is produced can markedly reduce or even and fully evaluated as to reasons for failure. Of course, eliminate dependence on narcotic analgesics.11 Patients with medical treatment must be individualized, but it generally cancer pain are also more likely to have good results from consists of a combination of analgesics, antidepressants, and ablative procedures because their pain derives from direct often anticonvulsants. Nonsurgical peripherally applied stimulation of pain receptors and is transmitted in pain neuroaugmentative and counterirritation techniques such pathways (i.e., nociceptive), which are eliminated by neu- as transcutaneous electrical nerve stimulation (TENS) and rodestructive procedures.7 These patients are also often un- acupuncture also may be of benefit,4–6 and these should be likely to survive long enough for the benefits of the proce- tried if indicated. For the purposes of this chapter, we shall dures to disappear or for new pain syndromes to develop. assume that best medical therapies have been instituted for Rehabilitation techniques can be of value in the manage- all conditions unless otherwise stated. ment of patients with cancer pain. TENS and acupuncture Neurosurgical options in the relief of chronic pain in- techniques have been used successfully for a variety of clude (1) the electrical stimulation of the brain, spinal cord, painful conditions associated with malignancy.12,13 Physical or peripheral nerves, (2) the implantation of devices to de- therapy with the use of joint manipulation, stretching exer- liver drugs to the CSF, and (3) the destruction or incision of cises, muscle strengthening, and massage can help control neural tissue.7 Many of the ablative neurosurgical options pain caused by contractures and other soft-tissue and joint for chronic pain treatment are depicted in Figure 50–1. complications due to inactivity.14,15 The application of heat and cold can also be used to treat various cancer pain Malignant Pain problems.12,14 Relaxation therapy also can be of benefit in improving pain control as well as general patient well- Pain is experienced by one third of cancer patients who re- being.13,14 By whatever means, the primary goal of treatment ceive active treatment and by three quarters of those with in such patients is to render them pain free for the duration advanced disease.8 Because cancer pain usually is associ- of their remaining life. Their restricted life span often makes ated with structural pathology, definitions of the extent of long-term rehabilitation planning unfeasible. Expecting pa- the disease and the nature of the specific underlying cause tients with malignant disease to attempt rigorous and ex- of the pain are essential in the assessment of each patient. pensive long-term programs to relieve pain may be unrea- Because cancer pain is always at least in part nociceptive sonable. The goal for the pain physician in such cases is to in origin, narcotic analgesics are the mainstay of pain management. About 90% of patients can obtain adequate pain relief with oral opioid analgesics.8 For most patients with medically intractable pain from malignancy, intrathecal opioids can provide relief in up to 90% of cases.3,9 For patients with SPECIAL CONSIDERATION lower-body, pelvic, and lower-extremity pain, intraspinal When considering neuroablative surgery for the morphine (either epidural or subarachnoid) is recommended. In patients with upper-body, cervical, or orofacial neoplasia, patient with cancer pain, the pain physician should intraventricular morphine can provide effective relief.10 Side consider (1) the patient’s ability to tolerate pain effects such as nausea and vomiting will occur in some pa- medication, (2) whether the patient would benefit tients. Although most patients can be effectively relieved more from rehabilitative procedures, and (3) the with antiemetics, some will not tolerate the dosage of in- patient’s life expectancy. trathecal morphine required for adequate analgesia. In pa- ABLATIVE NEUROSURGICAL TECHNIQUES 635

Figure 50–1 Ablative neurosurgical procedures used to control or alleviate chronic pain. relieve pain rapidly without significantly reducing the pa- signs, which are the standard for pharmacologic trials.16 Fre- tient’s quality of life. quently, neurosurgical centers report retrospective analyses with small numbers of differing patients treated with vari- Nonmalignant Pain ous modalities. Essentially all procedures developed for surgical pain management have therefore not been sub- Chronic pain from nonmalignant processes remains the jected to rigorous scientific randomized controlled trials. As major challenge to multidisciplinary pain physicians.16 Even a result, considerable uncertainty across the medical and among pain specialists, no consensus exists for the treat- neurosurgical communities remains as to the proper role of ment of many of the chronic nonmalignant pain syn- many neurosurgical procedures for the treatment of chronic dromes.17 The diversity of pain syndromes experienced by nonmalignant pain.18 patients and the relative rarity of neuroablative procedures Chronic benign pain can result from a variety of causes in this patient population make assessment of the relative and can be divided broadly into predominantly nociceptive, merits or otherwise of the various procedures difficult to neuropathic, psychopathologic, or the most common sub- quantitate.18 By their nature, surgical procedures are not set, a combination of all three. Intractable nociceptive amenable to double-blinded, placebo-controlled study de- pain results from nonneural tissue damage from trauma, 636 SURGICAL PROCEDURES inflammation, or degenerative or visceral disease processes. these patients. One possible exception is facet joint denerva- Predominantly nociceptive pain, by its activation of periph- tion, a procedure still performed in some centers. For low eral pain pathways, implies that conventional medical or back pain and lower-extremity pain from FBSS or degenera- surgical therapy should be effective. To this end, rigorous tive disease, spinal cord stimulation is our recommended assessment and investigation for reversible causes should management. If this fails, a trial of oral opioids should be at- be performed. The increased sophistication of investiga- tempted; failing this, treatment with intraspinal narcotics tional techniques such as computed tomography (CT) and should be considered. magnetic resonance imaging (MRI) as well as electromyog- This chapter does not discuss trigeminal neuralgia, al- raphy (EMG) has led to the improved understanding of though it does mention the treatment of anesthesia dolorosa many pain syndromes [e.g., failed-back syndrome, neural- following ablative procedures for this diagnosis. Although gias, and post spinal cord injury (SCI) pain], often allowing the occasional trigeminal and glossopharyngeal neuralgia more definitive conventional treatment; however, many pa- patient may require partial or complete rhizolysis, these are tients still fail to respond to conventional treatment regimes. rare exceptions. Management of trigeminal neuralgia is the Psychological and psychosocial, as well as physical, factors subject of Chapters 21–23. often confound the accurate assessment of chronic pain patients. For this reason, as an initial strategy, most of such complex patients should be referred initially to multidisci- SPECIAL CONSIDERATION plinary pain clinics, where all aspects of the patient can be assessed and managed. Except in specific circumstances, neuroablative Chronic neuropathic, or deafferentation, pain results techniques generally are not recommended for from neural tissue damage from a wide variety of causes chronic nonmalignant pain because (1) the mech- throughout the nervous system. More commonly, it occurs anisms, pathway, and subcortical integration of in the peripheral nervous system following some form of ac- cidental or surgical trauma. It also may follow ischemic or such pain are not yet understood; and (2) these hemorrhagic vascular disease in the CNS. Typically, neuro- procedures often fail to provide long-term pain pathic pain is diffuse and burning or stinging and is superfi- control. cial rather than deep in location. It tends to be constant and is often triggered by cutaneous stimuli. In general, unless the patient has prominent allodynia or hyperpathia, neuropathic pain will respond better to stimulation techniques or NEUROABLATIVE PROCEDURES to the administration of intrathecal opioids, with neurode- The various ablative procedures and their indications and structive procedures held in reserve if these fail.19 results, commencing with the peripheral nervous system Frequently, both nociceptive pain and neuropathic pain and then moving centrally in first the segmental and then are present simultaneously. Pain syndromes following SCI the suprasegmental CNS are discussed in the following illustrate this point. A high percentage of patients experi- sections. ence chronic pain following SCI.20 Pains may be of different types. Initial local pain at the site of injury is due to inflammation and damage to local structures with associated mus- Peripheral Neurectomy cle spasm. More chronic midline pain can result from spinal instability and muscle spasm or occasionally the develop- Neurectomy involves the resection of a portion or the whole ment of a posttraumatic syrinx. Radicular pain occurring at of one or more cranial or peripheral nerves. Ambroise Pare the level of the injury is often due to nerve-root injury or im- is credited with performing the first neurectomies in the late pingement, and management is directed at decompression sixteenth century. He recognized that sectioning nerves in or stabilization of the affected region. Pains occurring below the arm could relieve chronic pain secondary to penetrating the lesion, usually described as central pain, spinal cord wounds. Despite its long history, however, neurectomy has dysesthesias, or deafferentation pain, consist of diffuse per- only a limited role in chronic pain management today.24 sistent burning or tingling sensations that resemble neuro- The advantage of peripheral nerve resection is its relative pathic pains elsewhere in the body.21 The exact mecha- surgical simplicity. Exposure usually can be performed with nism(s) of pain generation in this situation remain unclear.21 the patient under local anesthesia. Division of a major prox- Because of the complexity and varied causes in such pa- imal nerve leads to complete anesthesia in the territory of tients, pain following SCI is one of the most difficult to eval- the nerve, although neurectomy of distal sensory nerves uate and treat.22,23 usually produces only hypalgesia. The initial disadvantage Degenerative spinal conditions constitute a large sector of of proximal neurectomy is the associated loss of useful func- patients with chronic nonmalignant pain seen in pain clinics. tion in the region of the nerve. Motor and proprioceptive With the application of neurostimulation techniques and the loss can be troublesome to the patient, as can the complete use of implantable opioid pumps, we believe there is no role anesthesia produced, which can lead to repetitive trauma in for neuroablative surgery in these patients. Ablative proce- extremities and to decubiti and the development of Charcot dures in the past have been used for patients with “failed joints.24 Local anesthetic blockade preceding nerve section- back surgery syndrome” (FBSS) and chronic back pain; how- ing always should be performed to demonstrate to the pa- ever, we would strongly discourage any such measure for tient the expected postoperative deficit. ABLATIVE NEUROSURGICAL TECHNIQUES 637

The major drawbacks to peripheral neurectomy are its currence and deafferentation pain. Postherpetic neuralgia long-term problems. Adjacent intact sensory nerves sprout does not respond to neurectomy because the site of pathol- into the denervated area, gradually decreasing the size of ogy lies proximal at the dorsal root ganglion and spinal the anesthetic region. Typically, with the gradual return of cord. For body-wall pain, ganglionectomy or cordotomy is normal sensation over months to years, the original pain re- far more effective, and these procedures have supplanted turns. A new pain syndrome attributed to denervation hy- the use of thoracic neurectomy. persensitivity also may develop. This is the development of constant neuropathic pains in the region of loss of sensibil- ity.25 The pain initially may be felt arising between the nor- POINT/COUNTERPOINT mal and the insensate regions, the so-called border zone of pain. These pain syndromes can be severe, and little, if any, Peripheral neurectomy is both a useful and a sim- therapy is effective. These pains can be more difficult to ple procedure . . . manage than the original pain syndrome. Painful traumatic neuromas also may develop after the transection of periph- . . . but pain recurrence, neuropathic pain in the eral nerves, and these can mimic deafferentation pain.24 Because of the high incidence of pain recurrence, indica- region of the loss of sense, and the development tions for peripheral neurectomies in the trunk and extremi- of neuromas may become long-term problems. ties are uncommon. Painful malignancies in extremities today are managed better by localized treatment regimens rather than by more proximal neurectomy. Pain attributable Dorsal Rhizotomy and Dorsal to head and neck neoplasms is managed better with intra- Root Ganglionectomy ventricular morphine or more central procedures directed at the brainstem and upper cervical spine. Spinal dorsal rhizotomy has been used in the treatment of Following peripheral nerve injuries, the development of selected cases of cancer and nonneoplastic pain. First de- painful neuromas occurs in the minority of patients. Con- scribed in 1889, the popularity of the procedure has varied servative nonsurgical care for this select group can be diffi- as newer techniques have become available and as the com- cult.26 Neuroma excision, neurectomy, and nerve release plications and long-term results of rhizotomy have been as- procedures can yield substantial subjective improvement in sessed. Today, the use of implantable pump systems for some patients.26 In patients with distal sensory neuromas, epidural or intrathecal opiate delivery and the use of neu- about 40% will have a greater than 50% reduction in pain rostimulation techniques have made the use of dorsal rhizo- scores following excision and proximal nerve implantation tomy for the management of pain essentially of historical in- into bone or muscle.26 For patients with in-continuity neuro- terest only. mas of major proximal sensorimotor nerves with significant Dorsal rhizotomy involves sectioning of the dorsal nerve residual distal function, neither neurectomy nor internal roots. Various modifications over the years have been em- neurolysis is effective alone; these are managed best by ployed in an effort to obtain greater pain relief while reduc- other methods, such as peripheral nerve stimulation.26 ing the extent of deafferentation and avoiding undue morbid- Proximal nerve resection in amputee patients with pain- ity. Extradural and percutaneous approaches29,30 have been ful stump neuromas can be of benefit; however, it is impor- described, while partial rhizotomy31 and the use of radio- tant in such patients to ascertain that the pain is clearly frequency lesions32 have also been used. Regardless of the caused by local pressure on the neuroma; in this case, the technique used, the procedure suffers serious drawbacks. presence of a positive Tinel’s sign is of considerable prog- The extensive denervation required for adequate analge- nostic significance. Nonspecific stump pains or phantom sia with rhizotomy requires multiple nerve-root sectioning, limb pains do not respond to nerve resection.24,27 and the subsequent loss of proprioception makes the proce- Rarely, patients with meralgia paresthetica will require dure inappropriate for extremity pain. The procedure is sectioning of the lateral femoral cutaneous nerve. Conserva- tolerated better in the thoracic and sacral regions, where tive measures such as the removal of restrictive clothing, greater success has been reported, particularly for malig- weight loss, restriction of activities involving thigh exten- nant pelvic and perineal pain.33 Extensive denervation of sion, and the administration of nonsteroidal antiinflamma- the pelvic region requires bilateral resection of the S2–S3 tory medication achieve relief in more than 90% of cases, and nerve roots for pain relief. The resulting deficits typically all should be attempted before contemplating any form of mandate that colostomy and urinary diversion procedures surgery.28 Local anesthetic blocks can provide long-lasting be performed prior to rhizotomy or ganglionectomy.34 relief as well as confirm the diagnosis.28 The most common As with peripheral neurectomy, pain recurrence is the surgical procedure used for meralgia paresthetica is decom- major reason for failure. Loeser reported good short-term pression alone, which avoids the potential risk of denerva- success (63%) in 45 patients; however, this had fallen to 28% tion pain. If decompression fails, neurectomy still can be success after 3 months.35 Onofrio and Campa reported 50 to used with good long-term results for recalcitrant patients. 80% failures in 286 patients, which led them to conclude Neurectomy has been used in the past to treat chest-wall that rhizotomy had only a “poor-to-fair chance of succeed- pain from malignancy, painful costochondral joints, the ing.”36 Early pain recurrence may be due to disease progres- “postthoracotomy” syndrome, and painful abdominal scars. sion or failure to section adequate nerve roots. As with pe- Long-term results have not been satisfactory because of re- ripheral neurectomy, late recurrences occur as the area of 638 SURGICAL PROCEDURES anesthesia reduces in size because of the sprouting of collat- modulation of the nociceptive afferents. The medial portion eral neural pathways in the peripheral and central nervous transmits the excitatory effects of each dorsal root to the ad- systems.24 The most feared long-term pain problem is the jacent segments, and its lateral part conveys the inhibitory development of dysesthetic or deafferentation pain. Initially influences of the substantia gelatinosa into the surrounding reported as a rare (4 to 5%) occurrence,37 recent experience neurons.44 Following avulsion injuries of the dorsal roots suggests a far greater incidence. Pagni et al38 reported anes- in either the brachial plexus or the lumbosacral (conus thesia dolorosa occurring in 8 of 15 (53%) patients after medullaris) region, a severe form of deafferentation pain long-term review. They now regard rhizotomy as high-risk may develop. The pain typically occurs immediately after in- surgery and no longer perform this operation. Other proce- jury, although it may be delayed in onset for up to 3 months. dures such as ganglionectomy or intraspinal morphine fol- It occurs in anywhere from 20 to 90% of patients suffering lowed by cordotomy or mesencephalotomy are preferred traumatic brachial plexus root avulsion and frequently fol- over spinal rhizotomy for patients. lows high-speed motor vehicle or motorcycle accidents. Dorsal rhizotomy originally was performed as a treatment for the relief of monoradicular sciatica. Initial successful results were reported; however, later series with longer SPECIAL CONSIDERATION follow-up highlighted the shortcomings of the procedure for nonmalignant pain syndromes. Apart from the develop- Lumbosacral plexus avulsion with resultant ment of dysesthetic pains, the most common drawback was lower-extremity deafferentation pain is being en- the return of the original pain complaint. Numerous investi- countered increasingly as more patients with ex- gations have sought an anatomic explanation for this complication. Afferent fascicles bridging adjacent rootlets and tensive pelvic or sacral fractures survive their ini- the conveying of afferent transmission by the sympathetic tial trauma. rami were proposed, leading to variations in the technique used to divide such connections. These efforts met with little or no success. The exact mechanism of pain generation is poorly under- The finding that up to 30% of all axons in the ventral root stood. Following injury to the DREZ, the balance between ex- were unmyelinated afferent fibers derived from the dorsal citatory and inhibitory mechanisms is altered. A reorganiza- root ganglion39–41 in contravention of the law of Bell and tion occurs in the dorsal horn, leading to an increased Magendie suggested the possibility that resection of the proportion of nociception-signalling neurons in the deeper dorsal root ganglion may be the procedure of choice in pa- layers of Rexed’s lamina. Numerous changes to the neuro- tients in whom rhizotomy was contemplated. The proce- transmitters and neuromodulators have been identified.44,45 dure was first performed in 1966 by Scoville.42 Initial enthu- Decreases in substance P, somatostatin, and metenkephalin as siasm was maintained with reports of high success rates for well as other transmitter substances all are thought to lead to pain relief. More recent studies have been less enthusiastic, downregulation of inhibitory neurotransmitters in the dorsal particularly for the treatment of sciatica, in which it was re- horn.45–48 This probably causes spontaneous high-frequency ported to be no more effective than dorsal rhizotomy, with discharges of dorsal horn neurons, which normally signal no- the return of the original pain syndrome. In our own experi- ciceptions at high firing rates (wide-dynamic-range neurons). ence using dorsal root ganglion resection to treat chest-wall The resultant hyperactivity of these second-order neurons pain, the postthoracotomy syndrome (intercostal neuralgia), leads to the perception of pain.45,47,49 The basis for performing and occipital neuralgia, we found this procedure has accom- a lesion in the dorsal root is to destroy this region of neuronal plished long-term pain relief. dysfunction, thereby reducing nociceptive input.50 Thoracic extradural ganglionectomy is a relatively safe The first DREZ lesioning procedure was by a DREZ procedure with relatively minimal associated morbidity. We myelotomy by Sindou in 1972 on a patient with chronic perform the procedure for indications in which a thoracic pain due to Pancoast’s syndrome.51 More recently, Nashold dorsal rhizotomy otherwise would be contemplated. We be- and colleagues performed DREZ lesioning using radiofre- lieve dorsal root ganglionectomy is not appropriate for sci- quency.52 Success with other patients with malignant pain atica, as recommended by some authors,43 and that aggres- syndromes encouraged further efforts of the treatment of sive rehabilitation should be attempted on all patients with chronic pain following brachial plexus avulsion, SCI, and degenerative spinal conditions before attempting any surgi- peripheral nerve injury.44,46,48 The procedure involves the cal options. If surgery becomes necessary, neurostimulation destruction of the lateral portion of the dorsal horn (where or intrathecal narcotics would be more appropriate than any C fibers predominate) over several segments. During the neurodestructive procedure. procedure, Lissauer’s tract and laminae I through V of Rexed are destroyed by knife, radiofrequency thermocoagu- Dorsal Root Entry Zone Lesions lation, or laser. By sparing the medial portion of the dorsal horn where large-diameter A fibers enter, an attempt is The dorsal root entry zone (DREZ) is an entity that includes made in Sindou’s technique to avoid complete abolition of the central portion of the dorsal root, the tract of Lissauer, tactile and proprioceptive sensation,44 which theoretically and layers I to V of the dorsal horn, where the afferent fibers may preserve some geometric antinociception. synapse with the origins of the sensory pathways.44 The tract Since its introduction, the DREZ procedure has been suc- of Lissauer plays an important role in the intersegmental cessful in treating pain caused by plexus avulsion, border- ABLATIVE NEUROSURGICAL TECHNIQUES 639 zone pain in SCI, and the paroxysmal pain associated with in which the primary pathology is avulsion of the dorsal peripheral nerve injuries such as amputation or herpes root.46,48,56 Only in select cases in which the preceding fea- zoster.3,44 It is also effective for topographically limited can- tures predominate would we recommend performing a cer pain such as Pancoast’s syndrome; however, for more DREZ lesion in such patients. extensive cancer, intrathecal morphine or high cervical Nucleus caudalis lesion was developed by Nashold as a spinothalamic cordotomy is preferable.44 tractotomy/nucleotomy (caudalis DREZ) extension of DREZ For patients in whom the dorsal root and dorsal horn are lesions to the trigeminothalamic system for the relief of fa- primarily affected by the injury, DREZ is the most beneficial cial pain syndromes.18 Two facial pain syndromes character- treatment. More than 85% of patients report good to excel- ized by deafferentation pain are trigeminal postherpetic lent pain relief following lesioning for painful brachial neuralgia (most commonly encountered in the ophthalmic plexus avulsion.3,48,51,53–55 Long-term follow-up of such pa- division) and facial anesthesia dolorosa. The latter occurs tients found continued relief after periods of 14 months to after traumatic injury to the divisions or branches of the 10 years.3 In lesions for root avulsions of the conus med- trigeminal nerve, usually following destructive procedures ullaris, 54% of 39 patients remained pain free with no med- for trigeminal neuralgia.3 Patients with anesthesia dolorosa ication, whereas an additional 20% required only simple following percutaneous rhizotomy or open rhizolysis or analgesics, with an average follow-up of 3 years.56 with postherpetic neuralgia constitute a considerable prob- In a recent series using microsurgical techniques, compli- lem in management because they generally experience se- cations were few, with only transient and nondisabling sen- vere constant burning sensations with shocklike exacerba- sory deficits and mild weakness reported.3 In procedures for tions. The pains are generally refractory to standard narcotic conus and cauda equina trauma, bladder and sexual dys- therapy. function occur in less than 10% of patients.56 Mortality is Lesions placed at the spinomedullary junction in the dor- also low in recent series. Only 2 deaths in a series of 600 solateral sulcus up to the level of the obex interrupt the ipsi- were reported by Nashold, although in a series of bedridden lateral nociceptive secondary afferent neurons from the patients death occurred in 5 of 53 patients.3,48 spinal trigeminal nucleus caudalis. Other orofacial primary Patients with chronic deafferentation pain from plexus nociceptive afferents entering the brainstem from either the avulsion present a considerable challenge to pain-manage- 7th, 9th, or 10th cranial nerves all descend in this compact ment professionals. The pain is usually refractory to virtually tract in the dorsolateral medulla.58 all conventional therapies. Standard analgesics and narcotics Initial anatomic studies on the course of the trigeminal are usually ineffective, and medical management has been nerve fibers in the medulla were performed by Sjoquist limited essentially to antidepressant, sedative, and anticon- in 1938.59 With the realization that nociceptive fibers de- vulsant medications combined with psychological counsel- scended farther than tactile afferents, the objective of sup- ing.57 An initial trial of conservative therapy should be at- plying analgesia without anesthesia could be attempted. tempted in all cases because in rare instances the pain will Early attempts at open trigeminal tractotomy and nucleo- gradually resolve. If the pain continues longer than 3 months, tomy met with high morbidity and mortality rates.3,60 The it is almost always permanent. In such cases, little is to be first description of performing nucleus caudalis radiofre- gained from lengthy rehabilitation programs. The unrelent- quency lesions was in 1972 for postherpetic facial pain.61 ing nature of the pain itself usually severely limits the pa- Percutaneous procedures aimed at the caudalis nucleus also tient’s activities. Given the reliable results obtained by per- were attempted; however, long-term experience and high forming microsurgical DREZotomy for plexus avulsion pain, morbidity rates discouraged many surgeons and pain we recommend proceeding directly to a DREZ procedure physicians.3 The results of these procedures still are poorly rather than continuing with more conservative therapies. reported, with essentially only one large series available for In SCI, more than 75% pain reduction can be reliably review. In 1987, the method of performing nucleus caudalis achieved for the radicular burning pains at the level of the DREZ was first described.59 With modern anesthesia and injury or for those patients with a preponderance of allody- with the use of microsurgical techniques and new insulated nia or electric-shock-like border-zone pains. Pains in the re- electrodes, the procedure is safer than in previous iterations. gion below the injury site, in particular in totally anesthetic The operation frequently causes ipsilateral upper limb regions in the perineum or sacrum, have not been relieved ataxia; however, this generally improves with time.3 Great reliably by the DREZ procedure.44 As mentioned previously, care needs to be taken with these patients because they are rigorous assessment of each patient is essential because of generally elderly and infirm; however, with meticulous at- the multiplicity of causes for SCI pain. For patients with dif- tention to detail and careful patient selection, significant fuse pain below the level of injury, rehabilitation is the benefits can be attained. The results of caudalis DREZ are in major mode of treatment; however, in a minority of DREZ no way comparable to that obtained with plexus avulsion, lesion cases, ablative techniques can offer significant im- and fewer than 50% of patients will obtain relief. Topical provements in pain control. cold packs and occasionally heat applications can provide In peripheral nerve injury pains, such as postamputation good short-term relief for some patients. For patients with- pain and postherpetic neuralgia, significant relief from out prominent allodynia or hyperpathia, trigeminal gan- DREZ lesions appears to be obtained only with the paroxys- glion or deep-brain stimulation is preferred.3 For patients mal components (electrical shooting pains) or pains trig- with severe trigeminal deafferentation pain refractory to gered by superficial stimulation.44 Most experience suggests oral analgesics, we recommend that open caudalis DREZ that peripheral pain states do less well than pain syndromes surgery be attempted for pain relief. 640 SURGICAL PROCEDURES

from a variety of causes that may or may not include direct POINT/COUNTERPOINT trauma to a nerve trunk. These include trauma such as lacer- ations, sprains, and fractures of the hands, feet, or wrist, and DREZ lesioning has proven highly successful in multiple medical conditions such as myocardial infarction, patients suffering from plexus avulsion pain . . . diabetic neuropathy, multiple sclerosis, or stroke. There is little evidence of the “reflexive nature” of RSD. In some cases, . . . but it has been less successful for patients the role of the sympathetic nervous system is questionable with peripheral nerve injury pain. and “dystrophy” may or may not be part of the syndrome. In an attempt to clarify terminology and eliminate the physiologic bias that the term RSD implies, the term CRPS has been devised to describe these syndromes. CRPS I is essentially Sympathectomy RSD with no clear nervous system injury; CRPS II was previously categorized as causalgia. Sympathectomies were first performed in the 1890s for the In the past, some authors reported a high rate of sponta- treatment of diverse conditions such as epilepsy, exoph- neous remission over periods of several weeks to many thalmic goiter, spasticity, and angina pectoris.62 Surgical tech- years.69,70 Although some patients do experience remission, niques and indications were refined over the next decades to in most the pain persists and often progressively worsens. include peripheral vascular disorders, hypertension, and Narcotic addiction and psychiatric problems rapidly de- pain. Whereas the advent of effective antihypertensive ther- velop in such patients because of the unrelenting nature of apy has made the use of sympathectomy obsolete for blood the pain.69,71,72 Trophic changes, which may be a source of pressure management, it remains an effective technique for pain, develop rapidly and over time become irreversible.69 It the management of sympathetically maintained pain and in is imperative that early, proper management of causalgia some cases of ischemic vascular disease.63–65 (CRPS II) be instituted. Little is gained with lengthy trials of Tissue damage in the extremities with and without obvi- conservative therapy if initial improvement is unsatisfactory. ous nerve injury sometimes is followed by diffuse burning Initial management includes physical therapy, pharma- pain and hyperalgesia, which can be relieved by blockade of cologic therapy, psychotherapy, and various anesthetic sym- the efferent sympathetic activity to the extremity. Sponta- pathetic blocks.62 Physical therapy and topical applications neous pain and hyperalgesia also may be associated with al- of cold and heat can provide good symptomatic relief in the terations to blood flow following either vessel occlusion or early phases of the pain. Active joint manipulation is impor- vasoconstriction. tant to ensure that contractures and trophic changes do not Controversy in the literature abounds regarding pain syn- impair function. The application of TENS also has been dromes that are responsive to sympathectomy. Terms such as used early and has been successful in alleviating discom- causalgia both major and minor (usually following nerve in- fort.1 Anesthetic blocks confirm the diagnosis and also can sult), reflex sympathetic dystrophy, Sudek’s atrophy, and provide relief from the pain for varied lengths of time, Raynaud’s syndrome all have been attributed to sympatheti- sometimes permanently. Accurate placement of the gua- cally mediated disorders with wide overlap of clinical situa- nethidine block using radiographic guidance is essential to tions and underlying causes. Typical features of sympatheti- ensure diagnostic accuracy. If improvement is prolonged, cally mediated pain disorders include vasospasticity and therapy should be repeated as needed.62,69 dystrophic features in which the joints and skin are affected Permanent or semipermanent interruption of sympa- more than other tissues. If dystrophic features predominate thetic activity can be produced by interrupting the par- over pain, the condition is termed Sudek’s atrophy. avertebral sympathetic ganglion chain. Dividing the sym- In an effort to reduce confusion, the term sympathetically pathetic chain below the first thoracic ganglion with maintained, or sympathetically mediated, pain has been used to resection of the T-2 sympathetic ganglion will interrupt encompass a spectrum of conditions that have in common sympathetic supply to the upper extremity. This now can the feature that the pain can be relieved by interruption of be performed by either open or endoscopic means. Percu- sympathetic outflow to a body region,64,66–68 although when taneous ablation of the sympathetic chain and T-2 ganglion Mitchell first coined the term causalgia, he intended for it to can be by chemical, electrical, or radiofrequency tech- designate simply “burning pain.” Causalgia as a syndrome niques. Open surgical procedures have high reported has taken on significantly more meaning; it is now reserved success rates (85 to 100%) with relatively low (2.5 to 5%) for the syndrome of sustained diffuse burning pain, allody- complication rates.3,48,69,73 The results will vary greatly ac- nia, and hyperpathia following traumatic nerve lesions. It cording to the precise condition being treated and the tim- is often combined with vasomotor and sudomotor distur- ing of the procedure in the course of the disease.62 Pain bances and later trophic changes.69 In most cases, the injury syndromes are best treated early in the course of the dis- to the nerve is a high-velocity missile and is partial. Most ease before the development of trophic changes.62,64,67,68 It commonly affected are the median and sciatic nerves as well is therefore important that patients be referred early in the as the brachial plexus. The pain usually develops within 1 course of the disease to multidisciplinary pain clinics so week after the injury and almost always is described as that rapid diagnosis and early treatment regimens are in- burning in character. The pains are intense, diffuse, and con- stituted. Sympathetically maintained pain represents one stantly present with superimposed exacerbations. of the few pain syndromes in which early definitive treat- Reflex sympathetic dystrophy (RSD) has been an all- ment may include an ablative procedure rather than exten- inclusive term applied to similar pain syndromes that result sive medical or rehabilitation therapy. ABLATIVE NEUROSURGICAL TECHNIQUES 641

Treatment failures do occur, and the long-term recur- chest or body wall is best treated initially by intraspinal rence rates vary considerably in the literature between 0 and morphine.3 If pain escapes control with narcotic use, high 33%.62,65,66,69,74 By definition, initial treatment failure indi- cordotomy can be performed with a minimum of risk and cates a misdiagnosis, and therefore rigorous preoperative an acceptable degree of neurologic deficit. assessment is needed.18 Careful assessment of each patient Cordotomy has been used only rarely for nonmalignant following the procedure is important to ensure that com- pain syndromes because of concern over the development plete sympathectomy has been obtained. Clinical testing of postcordotomy dysesthesias or the return of the original using iodine starch to assess sweat production can be use- pain syndrome itself. Postcordotomy dysesthesias, the most ful. Repeated diagnostic blocks also can be used to ensure feared complication of cordotomy, is the development of that complete sympathetic denervation has been accom- causalgic or neuropathic type pains after a variable latent plished, and, if necessary, surgery should be repeated.2 period in the region rendered analgesic. As with the other Physical therapy is important in treating these patients forms of deafferentation pain following neuroablative pro- following pain relief to restore the patients to a fully func- cedures, postcordotomy dysesthias are typically diffuse tioning status. In the early phase of the disease process, burning sensations with superimposed lancinating exacer- however, they are often incapacitated by the pain itself. Be- bations. The incidence is reported variably as between 6 and cause of the excellent results obtained by surgery, we recom- 20%; however, data are difficult to interpret in many series mend that, after a suitable trial of mobilization, physiologic because of the variability of follow-up intervals.75 therapy, and sympathetic blockade, sympathetic ablation be considered as an early therapy. BRAINSTEM PROCEDURES Cordotomy FOR CHRONIC PAIN The goal of anterolateral cordotomy is to create a lesion in Patients with orofacial malignancies can survive for many the contralateral spinothalamic tract to interrupt a nocicep- years with modern oncologic treatment. Such patients may tive input caudal to the source of the pain. Because of the experience severe chronic pain as a result of involvement of multiple levels of entry of afferent fibers in the dorsal horn, the trigeminal, glossopharyngeal, vagus, nervus intermedius, a successful cordotomy will abolish only painful sensation or upper cervical nerves. These patients often experience se- on the contralateral side several dermatome levels caudal to vere pain on coughing, swallowing, or chewing. The use of the cordotomy site.58,75 It was first performed in 1912 by intraventricular morphine in such patients has dramatically Spiller and Martin using conventional laminectomy.76 For reduced the number of patients who will require open or several decades after this operation, cordotomy remained stereotactic destructive procedures; however, occasionally one of the most successful procedures for the management patients will be refractory to intrathecal opiates and will re- of chronic pain, especially for malignant processes.58 quire a selective destructive procedure. Rehabilitation ther- Despite the success of open cordotomy in relieving pain, apy has little to offer such patients apart from simple topical significant morbidity was inevitable, in part because of the treatments that may provide short-term relief. fragile nature of many patients undergoing the procedure. In Brainstem procedures for chronic pain can be divided into the mid-1960s, percutaneous techniques were developed in an three categories: (1) mesencephalotomies; (2) pontine, bulbar effort to make the procedure safer. Further refinement of per- and trigeminal tractotomies; and (3) the caudalis DREZ pro- cutaneous cordotomy with the use of myelography, radiofre- cedure. We have discussed the caudalis DREZ procedure. quency lesioning, intraoperative stimulation, and impedance Stereotactic trigeminal tractotomy and nucleotomy for the monitoring produced reports of high success rates (80 to 90% treatment of chronic pain caused by orofacial, or skull base or higher) and low morbidity and mortality rates.75,77,78 neoplasms have never achieved widespread acceptance; how- Initially, it seemed that the percutaneous methods would ever, mesencephalic spinothalamic and trigeminothalamic supplant the use of open cordotomy; however, this has not tractotomy (mesencephalotomy) are among the most effective occurred for a variety of reasons.58 Improved therapies for operations to relieve cancer pain of the face, head, neck, and many types of cancer have substantially reduced the number upper body. Between 80 and 100% of patients experience good of patients with intractable pain. Nondestructive alternatives or excellent levels of analgesia.3,79–81 Patients often achieve such as neurostimulation and opioid infusion pumps have dramatic relief of pain that does not correspond to the degree been successful in controlling pain in more than 90% of pa- of analgesia produced,3,81 which may be explained by the de- tients in whom pain is unrelieved by standard opioid regi- struction of the reticulothalamic projections from the peri- mens.13,75 The reduced number of patients being referred for aqueductal gray matter.3 Central neuropathic pain caused by cordotomy in recent years has led to fewer surgeons develop- the thalamic syndrome, trigeminal dysesthesias due to the lat- ing the necessary expertise to perform percutaneous proce- eral medullary plate syndrome, and, under some circum- dures with the ease and precision previously reported.75 In stances, phantom limb pain have been reported to respond addition, the secondary afferent fibers relating to pain may be satisfactorily to mesencephalotomy.60 so widely distributed and subject to so much anatomical vari- In the mesencephalon, a lesion placed at the medial ation that lasting relief is more likely attained only by section lemniscus interrupts the ascending spinothalamic and of the entire anterior quadrant. This is more confidently ac- trigeminothalamic tracts as well as the spinoreticular pain complished by an open cordotomy under direct vision.58 pathways with their extensive connections to the peri- Open and percutaneous cordotomy have been used aqueductal gray matter and reticular formation. The role of widely to treat malignant pain in the lower body. Pain in the the spinoreticular or paleospinothalamic tract is unclear, but 642 SURGICAL PROCEDURES it appears to provide suprasegmental reflex aversive re- tors for pain control. Initially, attention was directed toward sponses and transmission of diffuse, dull, and aching pain.82 the main ventrocaudal nucleus but soon shifted to the medial Mesencephalotomy for face and head pain interrupts the thalamus where results appeared to be better and compli- crossing trigeminothalamic and adjacent medial reticulo- cations fewer. In recent years, attention turned from lesion thalamic tract on the side opposite the patient’s pain. Its major making to chronic stimulation of the brain. We recommend indication is for nociceptive pain secondary to carcinoma of deep-brain stimulation in all cases of benign pain in which in- the head, neck, or brachial region. Unilateral or bilateral pain tracranial ablative procedures are being contemplated. can be treated. If pain is bilateral, the lesion is made contralat- Most experience has been in the management of cancer eral to the side of greatest pain. The relief obtained is probably pain, although thalamotomy and cingulotomy have been due to bilateral effects on the dorsomedial midbrain. The use used for chronic benign pain. Typically, the patient has been of MRI imaging and careful intraoperative physiological local- treated for widespread pain from disseminated malignan- ization either with evoked potentials or direct stimulation re- cies, such as occurs with multiple bony metastases. In gen- duced the operative mortality rate to about 5%.60 eral, with the exception of medial thalamotomy, all the in- Spiegel and Wycis were the first to use stereotactic tech- tracranial procedures appear to have greater success with niques for mesencephalotomies, reporting their experiences nociceptive rather than neuropathic pain. All these pro- in 1948.60 Open craniotomy for mesencephalic tractotomy cedures represent end-stage management and should be had been attempted previously but had fallen out of favor be- attempted only if all other treatment options have been ex- cause of the high rates of mortality, morbidity, and postopera- hausted. Targets for intracranial ablation include cingulo- tive dysesthesias. The latter has been recorded as high as 70% tomy, thalamotomy, hypophysectomy, and pulvinotomy. for open procedures compared with rates of around 15% for Experience with these procedures is limited primarily to a stereotactic series.83,84 Shieff and Nashold suggested that the few selected centers, each of which advocates a different occurrence of dysesthesias is related to the involvement of the procedure for similar patient groups. This makes full assess- medial lemniscus in cases of large lesions.60,83,84 By avoiding ment of the procedures difficult. All the procedures carry the medial lemniscus and the careful use of prelesioning the same risks of intracerebral hematoma, wound infection, stimulation, damage to this structure can be avoided and the and damage to neighboring brain structures. incidence of mesencephalic dysesthesias can be reduced.60 The cingulate gyrus was a target for lesions as early as 1948, Changes in ocular motility occur in virtually every pa- when it was approached by open operation.86 The application tient in whom a mesencephalotomy is performed.60 Loss of of stereotaxis for pain relief was first performed in 1962 by upgaze is almost universal but is rarely noticed by the pa- Foltz and White.87 The mechanism for pain relief is unclear, al- tient. As such, it is not often a significant complication. though it presumably derives from interruption of the Papez Diplopia is rare, occurring in 1 to 5% of patients, but it usu- circuit of the limbic system.86 Cingulotomy is performed by ally resolves with time. Careful prelesioning stimulation using standard stereotactic techniques and can be performed should help to minimize this complication. Ipsilateral devi- under local anesthesia. Bilateral lesions are performed in the ation of the eyes, nystagmus retractorious, convergence nys- anterior aspect of the cingulate gyrus. This procedure is well tagmus, and pupillary asymmetry also can occur but usu- suited to terminally ill patients with widespread metastatic ally result or improve within weeks. By performing the disease of musculoskeletal origin.86 In a study of patients with lesion at the level of the inferior colliculus rather than the intractable cancer pain, 51% had moderate to complete pain re- superior colliculus, it has been claimed that the incidence of lief at 3 months. Cingulotomy also has been used for benign ocular side effects is reduced (83.3% down to 16.7%); how- pain syndromes, with 45% of patients with varied causes re- ever, the success rate of the procedure appears somewhat porting complete pain relief.88 We would recommend cingulo- lower (75% down to 58.3%).60,85 tomy only for the treatment of widespread visceral or bone pain in cancer patients for whom standard pain regimens have failed and who cannot tolerate intrathecal or intraventricular INTRACRANIAL ABLATIVE PROCEDURES morphine. We do not believe there are any proven indications for cingulotomy in nonmalignant pain syndromes. Intracranial ablative procedures for pain control were al- Medial thalamotomy, despite being the first target of in- most abandoned in functional neurosurgery. In previous tracranial ablation, is no longer recommended.19 The re- decades, the high complication rates of open operative pro- ported success rates are low, and it has been supplanted by cedures discouraged physicians and patients from pursuing other treatment regimens. For nociceptive pain not amen- intracranial options for pain relief except as an extreme last able to cordotomy, mesencephalic tractotomy is more effec- resort. Since the late 1980s, techniques for performing such tive; however, as mentioned, it poses a greater risk of oculo- procedures have improved significantly. Stereotactic tech- motor disturbance.19 niques using CT and MRI guidance avoid the need for ven- Severe generalized bone pain from metastatic disease triculography for accurate intracranial localization. As a caused by endocrine-sensitive carcinomas, such as breast or result, procedures now can be performed under local anes- prostate cancer, has been relieved by pituitary ablation in thesia and even can be performed on an outpatient basis. As 84% of patients in combined series; however, as mentioned, a result, in some centers, intracranial ablative procedures for intraspinal and intraventricular morphine delivery has sub- chronic pain are being reinvigorated. stantially reduced the number of patients referred for un- The first stereotactic pain procedure was a dorsomedial controlled pain.3 Modern stereotactic techniques of pituitary thalamotomy performed by Spiegal and Wycis in 1947. Subse- ablation, especially using radiosurgery, have made the need quently, many sites have been targeted by various investiga- for open operations, always potentially hazardous in such a ABLATIVE NEUROSURGICAL TECHNIQUES 643 frail population group, unnecessary. Pituitary replacement structive procedure is performed.3 All measures to support is required postoperatively, and diabetes insipidus is a spo- such patients with symptomatic relief, whether medical, re- radic occurrence. In most cases, these complications can be habilitative, or surgical, also should be used. In selected managed with little difficulty. The exact mechanism of pain cases, when oral and intrathecal opioids fail to provide re- relief is unknown. It is accepted that it neither affects the lief, selected ablative procedures can provide improved limbic system nor lessens psychological suffering as a pri- quality of life. In such cases, the neurosurgeon must select mary mechanism.88 It has been postulated to act by hor- the procedure with the most likelihood of supplying ade- monal, hypothalamic, and neurotransmitter release mecha- quate analgesia with acceptable risk for the duration of the nisms. Pain relief is almost immediate and seemingly does patient’s life. not correlate with tumor regression.88 In more than 200 re- For patients with medically intractable nonmalignant ported cases, excellent pain relief was reported in 45 to 65% pain, the use of neuroablative procedures is limited. The of all patients, and 75 to 85% were able to stop opioid in- consequence of ill-advised neuroablation may be that the take.92 Mean duration of effect seemed restricted to about 3 patient is left with a larger area of painful anesthesia that months, which both limits its effectiveness and argues for lasts for the remainder of the patient’s life.3 Only when all contemporary intrathecal opioid management regimes. other avenues of treatment are exhausted should ablative Pulvinotomy was first performed in 1966. The general in- procedures even be considered. Neurostimulation or intra- dications suggested are identical to those of cingulotomy thecal drug delivery should be attempted in most cases be- and hypothalamotomy (i.e., intractable cancer pain). There fore destructive operations are performed. In selected con- is some suggestion that the long-term pain relief is more ef- ditions, such as plexus avulsion pain, sympathetically fective following pulvinotomy and has been recommended maintained pain, border-zone pains in SCI, and some cases for patients with longer survival times of up to 18 months.88 of anesthesia dolorosa and postherpetic neuralgia, the results of neuroablation are predictable enough to suggest neuroablation as a primary treatment modality. In all other CONCLUSIONS conditions, medical or rehabilitative techniques almost cer- tainly will be more successful. Most patients with medically intractable cancer pain should be managed with intrathecal morphine before any neurode-

REFERENCES

1. Bonica JJ. General considerations of chronic pain. In: intractable cancer pain in 90 patients. Neurosurgery. Bonica JJ, ed. The Management of Pain. Philadelphia: Lea 1996;39:57–62. and Febiger; 1990. 10. Andersen PE, Cohen JI, Everts EC, Beder MD, Burchiel 2. International Association for the Study of Pain. Classifi- KJ. Intrathecal narcotics for relief of pain from head and cation of chronic pain: description of chronic pain syn- neck cancer. Arch Otolaryngol Head Neck Surg. 1991;117: dromes and definitions of pain states. Pain. 1986;(suppl 1277–1280. 3):51. 11. Poletti CE. Open cordotomy and medullary tractotomy. 3. Coffey RJ. Neurosurgical management of intractable In: Schmidek HH, Sweet WH, eds. Operative Neurosurgi- pain. In: Youmans JR, ed. Neurological Surgery. 4th ed. cal Techniques. 3rd ed. Philadelphia: WB Saunders; 1995: Philadelphia: WB Saunders; 1996. 1557–1571. 4. Woolf CJ, Thompson SWN. Segmental afferent fiber in- 12. Arbit E. Management of Cancer Related Pain. Mt Kisco, duced analgesia: transcutaneous electrical nerve stimu- NY: Futura; 1993. lation and vibration. In: Wall PD, Melzack R, eds. Text- 13. Patt RB. Cancer Pain. Philadelphia: JB Lippincott; 1993. book of Pain. 3rd ed. Edinburgh: Churchill Livingstone; 14. Doyle D, Hanks GWC, MacDonald N, eds. Oxford Text- 1994:884–896. book of Palliative Medicine. Oxford: Oxford University 5. Firebrace P, Hill S. Acupuncture: How It Works, How It Press; 1995. Cures. New Canaan, CT: Keats Publishing; 1994. 15. Basmajian JV. Manipulation, Traction and Massage. 3rd ed. 6. Fields HL, ed. Nonsurgical peripherally applied neu- Baltimore: Williams & Wilkins; 1985. roaugmentative and counterirritation techniques. In: 16. Max MB, Portenoy RK, Laska EM, eds. The Design of Core Curriculum for Professional Education in Pain. Seattle: Analgesic Trials. New York: Raven Press; 1991. IASP Press; 1995. 17. Davies HT, Crombie IK, Lonsdale M, Macrae WA. Con- 7. Loeser JD. Introduction: ablative neurosurgical opera- sensus and contention in the treatment of chronic nerve- tions. In: Bonica JJ, ed. The Management of Pain. Philadel- damage pain. Pain. 1991;47:191–196. phia: Lea and Febiger; 1990:2040–2043. 18. Consensus Conference on the Neurosurgical Management 8. Fields HL, ed. Cancer pain. In: Core Curriculum for Pro- of Pain. Meeting Report. Neurosurgery. 1994;34:756–761. fessional Education in Pain. Seattle: IASP Press; 1995. 19. Tasker RP. Stereotactic surgery. In: Wall PD, Melzack R, 9. Karavelis A, Foroglou G, Selviaridis P, Fountzilas G. In- eds. Textbook of Pain. 3rd ed. Edinburgh: Churchill Liv- traventricular administration of morphine for control of ingstone; 1994:884–896. 644 SURGICAL PROCEDURES

20. Burke DC, Woodward JM. Pain and phantom sensation 39. Coggleshell RE, Applebaum ML, Fazen M, et al. Un- in spinal paralysis. In: Vinken PJ, Bruyn GW, eds. Hand- myelinated axons in human ventral roots, a possible ex- book of Clinical Neurology. Vol. 2. Amsterdam: North- planation for the failure of dorsal rhizotomy to relieve Holland Publishing; 1976. pain. Brain. 1975;98:157–166. 21. Cioni B, Meglio M, Pentimalli L, Visocchi M. Spinal cord 40. Coggleshell RE. Afferent fibers in the ventral root. Neu- stimulation in the treatment of paraplegic pain. J Neuro- rosurgery. 1979;4:443–448. surg. 1995;82:35–39. 41. Hosobuchi Y. The majority of unmyelinated afferent 22. Davidoff G, Roth E, Guarracini M, et al. Function limit- axons in human ventral roots probably conduct pain. ing dysesthetic pain syndrome among traumatic spinal Pain. 1980;8:167–180. cord injury patients: a cross-sectional study. Pain. 1987; 42. Scoville WB. Extradural spinal sensory rhizotomy. J 29:39–48. Neurosurg. 1966;25:94–95. 23. Beric A, Dimitrijevic MR, Lindblom U. Central dyses- 43. Taub A, Robinson F, Taub E. Dorsal root ganglionectomy thesia syndrome in spinal cord injury patients. Pain. for intractable monoradicular sciatica. In: Schmidek HH, 1988;34:109–116. Sweet WH, eds. Operative Neurosurgical Techniques. 3rd 24. Loeser JD, Sweet WH, Tew JM, van Loveren H. Neuro- ed. Philadelphia: WB Saunders; 1995:1585–1594. surgical operations involving peripheral nerve. In: Bon- 44. Sindou MP. Microsurgical DREZotomy. In: Schmidek ica JJ, ed. The Management of Pain. 2nd ed. Philadelphia: HH, Sweet WH, eds. Operative Neurosurgical Techniques. Lea and Febiger; 1990:2044–2066. 3rd ed. Philadelphia: WB Saunders; 1995:1613–1622. 25. Pagni CA. Central pain and painful anesthesia. In: 45. Wilkinson HA. Radiofrequency percutaneous upper- Krayenbuhl H, Maspes PE, Sweet WH, eds. Pain, Its thoracic sympathectomy: technique and review of indi- Neurosurgical Management. Part II: Central Procedures. cations. N Engl J Med. 1984;311:34–36. Basel: Karger; 1977:132–257. 46. Rossitch E, Abdulhak M, Ovelmen-Levitt J, Levitt M, 26. Burchiel KJ, Johans TJ, Ochoa J. The surgical treatment Nasnold BS Jr. The expression of deafferentation dyses- of painful traumatic neuromas. J Neurosurg. 1993;78: thesias reduced by dorsal root entry zone lesions in the 714–719. rat. J Neurosurg. 1993;78:598–602. 27. Sunderland S. Nerves and Nerve Injuries. Edinburgh: 47. Loeser JD, Ward AA Jr. Some effects of deafferentation Churchill Livingstone; 1972:486–503. on neurons of the cat spinal cord. Arch Neurol. 1967;17: 28. Williams PH, Trzil KP. Management of meralgia pares- 629–636. thetica. J Neurosurg. 1991;74:76–80. 48. Nashold JRB, Nashold BS. Microsurgical DREZotomy 29. Doubloon D. Root surgery. In: Wall PD, Melzack R, eds. in treatment of deafferentation pain. In: Schmidek HH, Textbook of Pain. 3rd ed. Edinburgh: Churchill Living- Sweet WH, eds. Operative Neurosurgical Techniques. 3rd stone; 1994:1055–1065. ed. Philadelphia: WB Saunders; 1995:1623–1636. 30. Uematsu S, Udvarhelyi GB, Benson DW, Siebens AA. 49. Blumenkopf B. Neuropharmacology of the dorsal root Percutaneous radiofrequency rhizotomy. Surg Neurol. entry zone. Neurosurgery. 1984;5:900–903. 1974;2:319–325. 50. Sjoquist O. Studies on pain conduction in the trigeminal 31. Sindou M, Fischer G, Goutelle A, Allegre GE. Microsur- nerve: A contribution to the surgical treatment of facial ACTA Psychiat Scand gical selective posterior rhizotomy. Pain. 1981;1(suppl): pain. . 1938;17:1–139. 354. 51. Sindou M. Etude de la Jonction Radiculo-medullaire Pos- terieure: La Radicellotomie Posterieure Selective dans la 32. Stolker RJ, Vervest ACM, Groen GJ. The treatment of Chirurgie de la Douleur [thesis]. Lyons, France: Univer- chronic thoracic segmental pain by radiofrequency percu- sity Claude-Bernard; 1972. taneous partial rhizotomy. J Neurosurg. 1994;80:986–992. 52. Nashold BS Jr, Urban B, Zorub DS. Phantom pain relief 33. Crue BL, Todd EM. A simplified technique of sacral rhi- by focal destruction of the substantia gelatinosa of zotomy for pelvic pain. J Neurosurg. 1964;21. Rolando. Adv Pain Res Ther. 1976;1:959–963. 34. Young RE. Dorsal rhizotomy and dorsal root gan- 53. Gorecki JP, Burt T, Wee A. Relief from chronic pelvic glionectomy. In: Youmans JR, ed. Neurological Surgery. pain through surgical lesions of the conus medullaris 4th ed. Philadelphia: WB Saunders; 1996:3442–3451. dorsal root entry zone. Stereotact Funct Neurosurg. 1992; 35. Loeser JD. Dorsal rhizotomy for relief of chronic pain. J 59:69–75. Neurosurg. 1972;36:745–750. 54. Campbell JN, Solomon CT, James CS. The Hopkins ex- 36. Onofrio BM, Campa HK. Evaluation of rhizotomy: review perience with lesions of the dorsal horn (Nashold’s op- of 12 years experience. J Neurosurg. 1972;36:751–755. eration) for pain for avulsion of the brachial plexus. 37. Sweet WH. Deafferentation pain after posterior rhizo- Appl Neurophysiol. 1988;51:198–205. tomy, trauma to a limb, and herpes zoster. Neurosurgery. 55. Friedman AH, Bullit E. Dorsal root entry zone lesions in 1984;15:928–932. the treatment of pain following brachial plexus avul- 38. Pagni CA, Lanotte M, Canavero S. How frequent is sion, spinal cord injury and herpes zoster. Appl Neuro- anesthesia dolorosa following spinal posterior rhizo- physiol. 1988;51:170–174. tomy? A retrospective analysis of fifteen patients. Pain. 56. Sampson JH, Cashman RE, Nashold BS, Friedman AH. 1993;54:323–327. Dorsal root entry zone lesions for intractable pain after ABLATIVE NEUROSURGICAL TECHNIQUES 645

trauma to the conus medullaris and cauda equina. J 74. White JC, Silverstone B. Pain and related phenomena Neurosurg. 1995;82:28–34. including causalgia. In: Woodhall B, Beebe GW, eds. Pe- 57. Pagni CA. Central pain and painful anesthesia: patho- ripheral Nerve Regeneration. Medical Monograph. Wash- physiology and treatment of sensory deprivation syn- ington, DC: US Government Printing Office; 1956:311– dromes due to central and peripheral nervous system 347. lesions. Prog Neurol Surg. 1976;8:132–257. 75. Tasker RR. Percutaneous cordotomy. In Schmidek HH, 58. Poletti CE. Open cordotomy and medullary tractotomy. Sweet WH, eds. Operative Neurosurgical Techniques. 3rd In: Schmidek HH, Sweet WH, eds. Operative Neurosurgi- ed. Philadelphia: WB Saunders; 1995:1595–1612. cal Techniques. 3rd ed. Philadelphia: WB Saunders; 1995: 76. Spiller WG, Martin E. The treatment of persistent pain 1557–1571. of organic origin in the lower part of the body by divi- 59. Bernard EJ, Nashold BS, Caputi F, et al. Nucleus caudalis sion of the anterolateral column of the spinal cord. DREZ lesions for facial pain. Br J Neurosurg. 1987;1:81. JAMA. 1912;58:1489–1490. 60. Bullard DE, Nashold BS. Mesencephalotomy and other 77. Rosomoff HI. Bilateral percutaneous radiofrequency brainstem procedures for pain. In: Youmans JR, ed. cordotomy. J Neurosurg. 1969;31:41–46. Neurological Surgery. Philadelphia: WB Saunders; 1996: 78. Meglio M, Cioni B. The role of percutaneous cordotomy 3477–3488. in the treatment of chronic cancer pain. Acta Neurochir 61. Hitchcock ER, Schvarcz JR. Stereotaxic trigeminal trac- (Wien). 1981;59:111–121. totomy for post-herpetic facial pain. J Neurosurg. 1972; 79. Amano K, Kawabatake H, Tanikawa T, et al. Long-term 37:412–417. follow-up study of stereotactic rostral mesencephalic 62. Wilkinson HA. Sympathectomy for pain. In: Youmans reticulotomy in patients with intractable pain. Appl Neu- JR, ed. Neurological Surgery. 4th ed. Philadelphia: WB rophysiol. 1986;49:105–111. Saunders; 1996:3489–3499. 80. Frank F, Frank G, Gaist G, et al. Rostral stereotactic mes- 63. Dohn DF, Sava GM. Sympathectomy for vascular syn- encephalotomy in treatment of cancer pain: a survey of dromes and hyperhidrosis of the upper extremities. Clin 40 treated patients. Acta Neurochir Suppl. 1984;33:437. Neurosurg. 1978;25:637–650. 81. Nashold BS, Slaughter DG, Wilson WP, et al. Stereotac- 64. Gybels JM, Sweet WH. Neurosurgical Treatment of Persis- tic mesencephalotomy. Prog Neurol Surg. 1977;8:35. tent Pain. New York: Karger; 1984:1045–1061. 82. Willis WD. The origin and destination of pathways in- 65. Hardy RW. Surgery of the sympathetic nervous system. volved in pain transmission. In: Wall PD, Melzack R, In: Schmidek HH, Sweet WH, eds. Operative Neurosurgi- eds. Textbook of Pain. New York: Churchill Livingstone; cal Techniques: Indications, Methods and Results. Vol 2. 1985:88–99. New York: Grune and Stratton; 1982:1045–1061. 83. Shieff C, Nashold BS. Stereotactic mesencephalotomy. 66. Dawson DM, Katz M. Reflex sympathetic dystrophy. Neursurg Clin Am. 1990;1:825–839. Neurol Chron. 1993;8:1–6. 84. Amano K, Kawamura H, Tanikawa T, Kawabatake H, 67. Schwartzman RJ, McLellan TL. Reflex sympathetic dys- Iseki H, Taira T. Stereotactic mesencephalotomy for pain trophy: a review. Arch Neurol. 1987;44:555–561. relief: a plea for stereotactic surgery. Stereotact Funct 68. White JC, Sweet WH. Pain: Its Mechanisms and Neurosur- Neurosurg. 1992;59:25–32. gical Control. Springfield, IL: Charles C. Thomas; 1955. 85. Shieff C, Nashold BS. Stereotactic mesencephalic tracto- 69. Bonica JJ. Causalgia and other reflex sympathetic dys- tomy for the relief of thalamic pain. Br J Neurosurg. trophies. In: Bonica JJ, ed. The Management of Pain. Phila- 1987;1:305–310. delphia: Lea and Febiger; 1990. 86. Hassenbusch SI. Intracranial ablative procedures for 70. Bingham JAW. Causalgia of the face: two cases success- pain. In: Youmans JR, ed. Neurological Surgery. 4th ed. fully treated by sympathectomy. BMJ. 1947;1:804. Philadelphia: WB Saunders; 1996:3541–3551. 71. Mitchell SW. Injuries of Nerves and Their Consequences. 87. Foltz EL, White LE. Pain relief by frontal cingulotomy. J London: Smith Elder; 1872. Neurosurg. 1962;19:89–100. 72. Ochoa J. The newly recognized painful ABC Syndrome: 88. Hurt RW, Ballantine HT. Stereotactic anterior cingulate thermographic aspects. Thermology. 1986;2:65. lesions for persistent pain: a report on 68 cases. Clin Neurosurg. 1974;21:334–351. 73. Robertson DP, Simpson RK, Rose JE, Garza JS. Video- assisted endoscopic thoracic ganglionectomy. J Neuro- surg. 1993;79:238–240.

COMMENTARY Neurosurgeons always have been in the forefront of the DREZ operation. Unfortunately, over the past 15 years, we treatment of intractable pain. Over many years, the neuro- have seen the arrival of the new pain specialists, anesthesi- surgeon has devised highly specific and specialized treat- ologists as pain experts, cancer pain being treated exclu- ments, such as selective rhizotomy, percutaneous cordo- sively by the oncologist, as well as the use of often untried tomy, stereotactic brain lesions, and, more recently, the alternative medical treatments. For the most part, the trend 646 SURGICAL PROCEDURES has been positive in that more attention has been focused on tomy when they leave their training. With the recent revival the specific problem of the patient in pain. The unfortunate of stereotactic procedures, there seems to be little emphasis spinoff is that many of the nonsurgical treatments currently on the treatment of pain. At a recent neurosurgical meeting, in use deny the patient a specific neurosurgical treatment, a patient was presented with head and neck cancer and in- which might give years of relief without drugs. A case in tractable pain that was successfully treated by using the point is the intractable pain of brachial plexus avulsion gamma knife technique to produce a midbrain lesion. which, before the DREZ operation, was untreatable. The pa- The DREZ has made it possible to give long-term pain re- tient with an avulsion is often young and can look forward lief of avulsions and paraplegia, and yet many neurosur- only to a long life of drugs and psychological deterioration geons are not familiar with the surgical technique, which re- and, at times, suicide. A DREZ operation done early will quires the use of a small commercially available electrode give at least 10 years of pain relief without drugs. and a radiofrequency lesion generator, both of which are Most pain specialists think of neurosurgery as a last re- found in most operating rooms. DREZ lesions are made in sort or do not consider it at all. The relief of patients with the region of the dorsal horn of the spinal cord, which is cancer pain is a good example of the neglect of the use of ex- readily seen under the operating microscope. The original cellent neurosurgical treatments. The neurosurgeon is called Sindou procedure was called a rhizidotomy and later was in only at the end, when there is little he or she can do. Nar- modified to the DREZotomy. We believe the important cotic drugs should be used only as an end treatment. The aspect of the DREZ operation is to destroy completely the morphine pump is overused in the early treatment of cancer dorsal horn and the secondary pain-afferent neuron, which pain, when a well-done percutaneous cordotomy could give become hyperactive after differentiation, although these hy- months of relief without pain. There is nothing worse than peractive neurons have been recorded both in the human the pain from head and neck cancer, and yet most cancer spinal cord and in the midbrain of patients with trigeminal pain specialists have never heard of a stereotactic mesen- pain. Despite these theories and human observations, the cephalotomy, which, done under local anesthesia, relieves pain enigma remains. I hope that in the future the neurosur- the patient until the end of his or her life. Example after ex- geon will reexamine this role in the treatment of intractable ample could be cited in which a specific neurosurgical oper- pain using the newer stereotactic and imaging technique ation could offer relief but is not used. Neurosurgeons bear and that they will become aggressive advocates for new some blame for the present state because few neurosurgical neurosurgical pain operations. residents are capable of doing a simple percutaneous cordo- Blaine S. Nashold Jr.