Browse the corpus

8. PAIN Konstantin V . Slavin istorically, surgery for pain has been a large part of general neurosurgical practice. A variety of destructive and decompressive interventions have been developed over the years, and a number of comprehensive textbooks have summarized neurosurgical involvement with management of all kinds of medically refractory pain syndromes. Over the past 40 to 50 years, the field of pain surgery has expanded tremendously. Introduction of non destructive options, primarily along the lines of electrical and chemical neuromodulation, has shifted the majority of care away from neurosurgeons and toward the continuously growing cohort of pain specialists, most of whom come from anesthesia, psychiatry, and neurology backgrounds. Nevertheless, surgical management of pain remains an important section of today’s neurosurgery. It is included in the core neurosurgical education curriculum and is an integral part of neurosurgical knowledge that is tested during the Oral Board Examination. Not surprisingly, cases involving complex pain conditions that require neurosurgical interventions routinely show up during examinations, and it is expected that examinees are comfortable performing these interventions and able to discuss indications, surgical details, outcomes, and complications. CASE 1 CLINICAL PRESENT A TION A 49- year- old woman presents with complaints of sharp and shooting pain in the left side of her face. The pain started 4  years ago and became progressively worse over time. It comes as a series of extremely painful attacks that last from a few seconds to about a minute, and between the attacks she is pain free. The neurological examination is intact; the areas that tend to trigger her pain are located in the upper lip and the gum of the upper jaw on the left side only; she is hesitant to brush her teeth, chew, and talk because each of these activities is associated with facial pain. She is currently being treated with oral gabapentin with only partial improvement of pain; previously, she had tried oral carbamazepine, which was very effective in controlling her pain but produced severe hyponatremia that prompted the change in medication regimen. DIAGNOSTIC WORKUP The first question is, obviously, what is the diagnosis? There are a number of tests that would help in establish ing it. A  correct answer would be trigeminal neuralgia (TN), and because it is a clinical diagnosis, the next steps in confirming it would be to obtain a detailed history and a standard imaging of the brain (preferably, magnetic reso nance imaging [MRI] of the brain with and without con trast, with emphasis on cerebellopontine angles), mainly to rule out secondary TN due to demyelinating disease (e.g., multiple sclerosis) or mass lesions (e.g., tumors, vascular malformations). TREA TMENT OPTIONS If the MRI does not show an underlying pathologic condi tion that would present with secondary TN, the classical clinical presentation and failure of medical management should be considered an indication for surgical treatment— and here the correct step would be to present the patient with all three main surgical options while discussing advantages and shortcomings of each of them. In a young and otherwise healthy person with classical TN, a procedure of choice would be microvascular decompression (MVD) of the trigeminal nerve through retromastoid craniectomy or craniotomy under general anesthesia.

with all three main surgical options while discussing advantages and shortcomings of each of them. In a young and otherwise healthy person with classical TN, a procedure of choice would be microvascular decompression (MVD) of the trigeminal nerve through retromastoid craniectomy or craniotomy under general anesthesia. Patients who are not medically fit to undergo open cranial surgery under general anesthesia (the age by itself is not a contraindica tion because there is a good amount of literature showing

with all three main surgical options while discussing advantages and shortcomings of each of them. In a young and otherwise healthy person with classical TN, a procedure of choice would be microvascular decompression (MVD) of the trigeminal nerve through retromastoid craniectomy or craniotomy under general anesthesia. Patients who are not medically fit to undergo open cranial surgery under general anesthesia (the age by itself is not a contraindica tion because there is a good amount of literature showing 76 • G OODMAN ’S N EUROSURGERY O RAL B OARD R E v IEW safety of MVD in elderly patients) may choose to proceed with percutaneous destructive surgery that may be done with radiofrequency (RF) thermocoagulation, glycerol gangliolysis, or balloon compression. Each of these is con sidered an appropriate intervention with consistently good results and expected partial sensory loss, with a general correlation between degree of sensory loss and duration of beneficial effects. RF retrogasserian thermocoagulation is considered the most selective intervention and is done with the patient awake because positioning of the electrode tip is determined by the patient’s feedback. Glycerol gan gliolysis may also be done in awake patients but requires much less patient participation. Balloon compression may be safely done under brief general anesthesia or deep seda tion because no patient cooperation is needed. Finally, an option of stereotactic radiosurgery (SRS) has to be brought up because this modality is an accepted way to treat TN— although it is generally less effective compared with other surgical interventions, at least in terms of the probability of making the patient pain free without medications and in terms of immediacy of pain relief. (The SRS effects may take between a few weeks to a few months to reach maximum, as opposed to open surgery and percutaneous destructions, which are expected to render the patient pain free by the end of surgery.) SRS appears to be most pre ferred by the patients because of its perceived lower inva siveness and minimal pain and discomfort experienced during the procedure. In cases in which the MRI of the brain shows demy elination, the choice of surgical intervention would shift away from MVD and more toward the percutaneous destruction or radiosurgery. Conversely, if one detects a mass lesion in the vicinity of the trigeminal nerve root, the surgical treatment should be focused on eliminating the lesion (surgery or SRS) rather than proceeding with treat ment of its symptom (TN). In radiosurgical treatment of tumors that present with TN, it is recommended to use standard tumor- related treatment protocols rather than a highly focused high- dose SRS, which is usually reserved for idiopathic TN or secondary TN due to multiple sclerosis. Detailed knowledge of the surgical procedure is expected during the Oral Board Examination— particular attention is generally paid to choice of anesthesia, positioning, approach, use of physiologic monitoring, and certain procedural nuances that may avoid certain complications. Complications are expected to happen no matter what, and it is very important to know what to expect, what can be done to minimize the risk, and most important, how to handle the complications when they occur. Microvascular Decompression MVD is considered a procedure of choice for classical TN due to its nondestructive nature and based on its consis tently shown best long- term success rate in terms of com pleteness of pain relief and incidence of recurrences. The procedure is done under general anesthesia; the patient is placed in a supine position with a shoulder roll or in a “park bench” lateral position with head turned horizontally so that the sagittal plane is parallel to the floor.

erm success rate in terms of com pleteness of pain relief and incidence of recurrences. The procedure is done under general anesthesia; the patient is placed in a supine position with a shoulder roll or in a “park bench” lateral position with head turned horizontally so that the sagittal plane is parallel to the floor. The incision is made behind the hairline— the goal of bone opening (with craniotomy or, more commonly, craniectomy) is to expose the junction of transverse and sigmoid sinuses and open the dura in such a way that the cerebellar hemisphere may be retracted medially and caudally to follow the superolateral corner of the posterior cranial fossa (the angle between the tentorium and petrous bone) to the arachnoid membrane of the cerebellopontine angle cistern. Multiple nuances are expected to be mentioned. First, it is recommended to use intraoperative physiologic moni toring, usually with continuous recording of bilateral brainstem auditory evoked responses (BAERs) and the electromyographic activity of the ipsilateral muscles inner vated by the trigeminal (masseter) and facial (frontal, peri orbital, and perioral) nerves. Second, it is recommended to wax the mastoid air cells that are exposed or violated during the drilling process. As a matter of fact, it is recommended to wax them twice— first during the opening and then again just before the closure. Next issue is the use of the micro scope. It is expected that the intradural part of the proce dure will be done under surgical microscope. Use of the endoscope to provide better visualization of neurovascu lar conflict may also be mentioned, but not using it would not be considered a departure from standard care because endoscope- assisted MVD so far is not universally accepted. Same with use of rigid retractor to expose the cerebello pontine angle: if the preoperative diuretics and spontane ous egress of cerebrospinal fluid (CSF) after opening the cerebellopontine angle cistern provide sufficient relaxation of the cerebellum, use of rigid retractor may be avoided. It is advisable to identify the cranial nerve VII- VIII complex that lies more inferiorly and more posteriorly to the cranial nerve V root in order to orient oneself. The greater petro sal vein of Dandy may be safely coagulated and cut; not doing this may result in hard- to- control venous bleeding if the uncut vein avulses from the tentorium in the middle of nerve root dissection (which is almost guaranteed to hap pen during the Oral Board Examination). In most cases, the offending vascular loop conflict will be easily identified. There may be one or two arterial vessels

sult in hard- to- control venous bleeding if the uncut vein avulses from the tentorium in the middle of nerve root dissection (which is almost guaranteed to hap pen during the Oral Board Examination). In most cases, the offending vascular loop conflict will be easily identified. There may be one or two arterial vessels PAIN • 77 that are compressing, indenting, and displacing the trigeminal nerve root a few millimeters away from the brainstem. Preoperative high- resolution imaging may help to visual ize the vascular structure before the operation. The most common offending vessel is the superior cerebellar artery, which is encountered in approximately 67.5% of patients; the second most common is the anterior inferior cerebel lar artery in 20% of cases; and in 25% of cases, a venous structure (such as an aberrant trigeminal vein) might be the cause of the conflict— this structure is visible on contrastenhanced T1- weighted MRI and not on magnetic reso nance angiography. After the vessel is dissected away from the nerve (not the other way around!), the decompression is completed by placing a nonabsorbable felt- like material between the nerve root and the vessel. The most commonly used material is shredded T eflon that is rolled into small patties or pledgets, which are inserted one after another to separate the vessel and nerve root. If the compression is venous (much less often), the vein is coagulated and cut. It is important to inspect the entire cisternal segment of the nerve from its entry into the brainstem all the way to the porus trigeminus because missed vascular compression may result in incomplete pain relief or early pain recurrence. In rare cases in which there is no identifiable vascular compression, one may use so- called internal neurolysis, combing the nerve root between its fascicles with a blunt hook or gen tly compressing the nerve (as was suggested in the 1950s, before MVD was introduced), rather than rushing to open transection of the nerve root (trigeminal rhizotomy). On completion of MVD, the dura is closed either pri marily or with a dural patch, and the bone defect is closed with autologous bone or cement with or without a plate. It may be helpful to mention multiple other steps that are routinely used (e.g., preoperative antibiotic, surgical timeout, use of Mayfield or similar head holder), but the examiners’ attention is usually paid to more specific surgical details. Complications The patient may develop CSF leak (particularly if the dural closure was not watertight), hearing loss (which will be partial if one forgets to wax the air cells and the patient develops muffled hearing from CSF in the mastoid cells, or complete if one does not use BAER monitoring and inadvertently stretches cranial nerve VIII during retraction), facial numbness due to excessive manipulation of the trigeminal nerve root, or incomplete pain relief after the MVD. Each of these complications has to be properly addressed as soon as it is discovered. CSF leak may require reexploration for patch ing the dural defect. Hearing loss would necessitate a con sultation with an ear, nose, and throat specialist but, just like facial numbness, is unlikely to require additional interven tion. Incomplete— or absent— pain relief may indicate that the compression was missed or the pledgets have moved; this may necessitate reexploration and re- decompression. However, if the pain returned more than 6  months after MVD, the risk for development of facial numbness from reexploration becomes so high that it may be prudent to consider other interventions rather than repeating MVD.

n was missed or the pledgets have moved; this may necessitate reexploration and re- decompression. However, if the pain returned more than 6  months after MVD, the risk for development of facial numbness from reexploration becomes so high that it may be prudent to consider other interventions rather than repeating MVD. Percutaneous Interventions It would be great to be equally familiar with all three destructive interventions that are commonly used for TN, but it is mandatory to know at least one of them in minor detail. The choice between RF thermocoagulation, glycerol gangliolysis, and balloon compression is usually dictated by the surgeon’s training and individual preferences, but in our practice we use the most selective intervention (RF thermocoagulation) whenever the patient is able to cooperate during the surgery and the pain involves either second or third, or both second and third, branch distribution. In patients with first branch involvement and in those with dementia or cognitive impairment or any other issues (including language barrier) that would prevent them from clearly com municating during the surgery, we prefer using the balloon compression approach, which does not require the patient’s cooperation and may be done under brief general anesthesia or deep sedation. The entry point and direction of the cannula are chosen based on H ärtel’s stereotactic technique of foramen ovale cannulation, with the skin entry placed 2 to 3 cm lateral to the corner of the mouth and 1 cm inferior to it. The can nula is aimed toward the ipsilateral medial epicanthus or slightly medial to the midpupillary line and to a point 1 to 3 cm in front of the tragus. The operator’s finger is placed in the patient’s mouth to prevent penetration of oral mucosa. W e routinely use a submental- vertex view of the patient’s skull through a standard C- arm fluoroscope. Some other centers use a C- arm direction that is parallel to H ärtel’s approach, tilting the C- arm until the foramen ovale is clearly visualized. The process of cannula insertion during RF procedures usually requires brief sedation; the patient is awakened when the cannula is in place, and the preganglionic nerve fibers are stimulated with low- voltage current to check the location of paresthesias and define the sensory thresh old. When the electrode is in the correct place (we usually use a curved electrode designed by T ew), the RF lesion is created by heating the electrode to 65 o to 70 o C for 60 to 90 seconds and then testing the patient for the expected

voltage current to check the location of paresthesias and define the sensory thresh old. When the electrode is in the correct place (we usually use a curved electrode designed by T ew), the RF lesion is created by heating the electrode to 65 o to 70 o C for 60 to 90 seconds and then testing the patient for the expected 78 • G OODMAN ’S N EUROSURGERY O RAL B OARD R E v IEW loss of sharp- dull discrimination and disappearance of trigger zones. The lesion is repeated in a different location if the painful regions are not covered by a single thermocoagula tion. A straight electrode may be advanced deeper to reach fibers of different branches; the curved electrode may be repositioned by retracting it into the cannula, rotating the cannula, and then redeploying the electrode in a different direction because an attempt to rotate the electrode itself would result in electrode fracture. When the loss of sharpdull discrimination is confirmed, the electrode and the cannula are removed. Balloon compression follows the same general approach— it may be worthwhile to move the entry point slightly higher, keeping it 2 to 3 cm lateral to the labial commissure to allow for more flat path for the balloon. The cannula is loaded with a sharp stylet to penetrate the skin; this sharp stylet is replaced with the blunt stylet when the can nula is 4 to 5 cm deep in the soft tissues. This is done to minimize risk for complications such as carotid artery puncture. After the stylet penetrates the dura of the foramen ovale, the cannula is lodged in the foramen, and then the stylet is replaced with the balloon, which is then advanced into Meckel’s cave under direct radiographic control with the C- arm in lateral projection. The balloon with a stylet in it is advanced until it is completely outside of the cannula and inside Meckel’s cave. When it is inflated with 1 mL of radiographic contrast (the contrast formulation should be safe for intrathecal application), the balloon would assume a typical pear- shaped appearance that indicates its position under the dural fold of porus trigeminus (Figure 8.1). This is the site where actual compression takes place, and during the balloon inflation it is not unusual to see a short period of bradycardia due to trigeminovagal reflex. The balloon is kept inflated for 90 seconds or so, although there are some published reports in which the balloon was inflated for a much longer time. The compression may be repeated if desired, and then the balloon and the cannula are removed. In case of balloon rupture, it may be replaced with another one— the contrast leaking into CSF as a result of such rupture is expected to be harmless. Complications Sensory loss after either of the percutaneous destructive trigeminal surgeries is not considered a complication; it is an expected effect of nerve destruction. The extreme numbness that is perceived as pain is called anesthesia dolorosa (AD)— and this is indeed a rare and hard- to- treat com plication. The best strategy to deal with AD is to avoid it, which is the rationale for not raising the temperature of the RF probe too high (above 85 o C) or lesioning for too long (longer than 90 seconds) and for not keeping the balloon inflated for longer than 3 minutes or inflating it with more than 1 mL of contrast. However, if AD does occur, the treatments are quite limited and not always effective. One would usually start AD treatment with tricyclic antidepressants (e.g., amitriptyline, nortriptyline) and use surgical interventions only if medical treatment fails. Both destructive sur geries (trigeminal tractotomy and nucleotomy, either open or computed tomography [CT] guided [Figure 8.2]) and Figure 8.1 Intraoperative fluoroscopic image during trigeminal balloon compression procedure.

amitriptyline, nortriptyline) and use surgical interventions only if medical treatment fails. Both destructive sur geries (trigeminal tractotomy and nucleotomy, either open or computed tomography [CT] guided [Figure 8.2]) and Figure 8.1 Intraoperative fluoroscopic image during trigeminal balloon compression procedure. Note the typical pear- shaped appearance of the balloon indicating the correct position of the balloon inside Meckel’s cave. Figure 8.2 Computed tomography– guided trigeminal tractotomy and nucleotomy. The contour of the uppermost spinal cord is delineated with intrathecal contrast; the needle is reaching the surface of the cord; and the electrode is in the posterolateral quadrant of the cord (the patient is positioned prone in the scanner).

Figure 8.2 Computed tomography– guided trigeminal tractotomy and nucleotomy. The contour of the uppermost spinal cord is delineated with intrathecal contrast; the needle is reaching the surface of the cord; and the electrode is in the posterolateral quadrant of the cord (the patient is positioned prone in the scanner). PAIN • 79 neuromodulation (motor cortex stimulation [Figure 8.3]) have been tried for AD with varying degrees of success. If the carotid artery is inadvertently penetrated during access to the foramen ovale, it is recommended to withdraw the cannula and abort the procedure. Because the carotid puncture is extracranial, the risk to the patient remains low, but it would be prudent to wait 1 or 2 weeks before attempting repeat trigeminal intervention. Stereotactic Radiosurgery Despite its relatively low effectiveness, SRS is an established modality for treatment of TN. Just like with everything else during the Oral Board Examination, it is advisable to avoid using the brand names, particularly because all currently available SRS devices (Gamma Knife and frame- based and frameless linear accelerators, including CyberKnife) have been reported to be used for TN treatment. The tar get of SRS is the cisternal segment of the trigeminal nerve root (Figure 8.4)— it is best visualized with special MRI sequences (e.g., FIEST A, CISS). If the patient cannot undergo MRI because of an implanted pacemaker or defi brillator, CT cisternography may be used for SRS target ing. The usual dose for the first SRS for TN would be 80 Gy; this number varies from center to center. As mentioned earlier, it takes some time for the SRS effect to reach maxi mum, which makes it different from other surgical inter ventions that work immediately. The results of SRS for TN are also less impressive than those of other approaches but are well received by the patients. CONTINGENCY PLANS If the surgical procedures do not work or if the pain recurs, the surgery may be repeated. It is not uncommon for TN patients to require more than one surgery, and there are no data to suggest that either SRS or percutaneous interven tions make subsequent MVD less effective, or vice versa. The challenge usually is to differentiate TN from trigeminal neuropathic pain (TNP) that is associated with constant pain and sensory deficit; even though TNP resembles a less typical presentation of TN (so- called TN type 2, in which constant pain dominates the clinical picture but trigger able shooting pains are also present). Recurrent TN, even in presence of neurological deficits, is an indication for surgery, but TNP is not expected to improve with either MVD or destructive interventions. Instead, TNP would be expected to respond to neuromodulation approaches such as peripheral nerve stimulation (PNS). CASE 2 CLINICAL PRESENT A TION A 68- year- old woman presents with severe pain in her hip, pelvis, and leg on the right side. The patient was diagnosed with metastatic cervical cancer that invaded her lumbar plexus on the right side. The tumor was deemed unresect able, and the patient underwent a course of radiation treatment and continues to receive systemic chemotherapy. Since the pain has become refractory to systemic opioid therapy, A B Figure 8.3 A: Functional magnetic resonance imaging (fMRI) indicating the location of hand and face motor activation in planning motor cortex stimulation for deafferentation facial pain. B: An epidural paddle- type electrode is placed over the mapped representation of the face in the contralateral motor cortex. In addition to fMRI- guided navigation, we use intraoperative recording of cortical activity for physiologic localization of the motor cortex. Note the mesh in retromastoid area from a previous intradural trigeminal rhizotomy procedure.

laced over the mapped representation of the face in the contralateral motor cortex. In addition to fMRI- guided navigation, we use intraoperative recording of cortical activity for physiologic localization of the motor cortex. Note the mesh in retromastoid area from a previous intradural trigeminal rhizotomy procedure. 80 • G OODMAN ’S N EUROSURGERY O RAL B OARD R E v IEW a neurosurgical service is being consulted for consideration of pain- relieving surgical interventions. The patient’s neurological examination is consistent with right lumbar plexopathy. She has a working colostomy and suffers from chronic urinary incontinence, most likely owing to focal invasion of pelvic structures. Her spine imaging does not show any neoplastic involvement of the vertebral column; there is no abnormal enhancement to suggest metastatic tumors on her brain imaging. She does, however, have metastases in her liver and lungs. SURGICAL DECISION MAKING In this clinical situation, the main question that should be asked by the surgeon is what is the patient’s life expectancy because this will determine the choice of surgical interven tion. As with most chronic pain conditions that require surgical intervention, there are two general approaches that are commonly used— neurodestruction and neuromodulation. The destruction is often irreversible, not testable with cer tainty, and not adjustable. Moreover, the results of destructive interventions are relatively short- lasting because pain would frequently recur in 6 to 12 months, possibly owing to inherent plasticity of the nervous system. Nevertheless, it remains attractive owing to its immediate onset of action and lower costs because an expensive implant is not needed for modulation. This choice between destruction and modulation comes up frequently when dealing with cancer- related pain because destructive interventions, such as cordotomy and myelot omy, may be done in a minimally invasive fashion and the short duration of the pain relief may be sufficient for many patients in terminal stages of malignancy. Electrical stimulation of the spinal cord and peripheral nerves may not be as effective because of the nature of the cancer pain: both SCS and PNS seem to work best for neuropathic pain condi tions, and cancer pain has a major nociceptive component due to actual injury of the tissues. T wo alternatives for highdose systemic opioids are intrathecal drug delivery (opioids, calcium channel blockers, various adjuvants including local anesthetics and adrenergic agents) and surgical interruption of pain- transmitting pathways— and for the patient pre sented here, the unilateral nature of the pain would make her a good candidate for a cordotomy procedure. If the patient’s life expectancy is longer than 6 months, the preference is given to intrathecal drug delivery, the modality that includes implantation of a programmable pump, commonly referred to as “morphine pump.” If, on the other hand, her life expectancy is 3  months or less, a cordotomy would be a more reasonable surgical choice. Intrathecal Morphine Pump The main rationale for intrathecal drug infusion is the ability to obtain good pain relief with significantly lower dose of the drug (so- called equianalgesic ratio) because the effect of 1 mg of intrathecal morphine is roughly equivalent to 100 mg of intravenous morphine. Such significant reduc tion of the medication dose translates into a better safety and satisfaction profile and allows one to avoid overse dation, constipation, and other dose- related opioid side effects.

ratio) because the effect of 1 mg of intrathecal morphine is roughly equivalent to 100 mg of intravenous morphine. Such significant reduc tion of the medication dose translates into a better safety and satisfaction profile and allows one to avoid overse dation, constipation, and other dose- related opioid side effects. Intrathecal delivery also allows for better compli ance, more steady effect without dosing- related fluctua tions, and with most recent pump models, an option of on- demand boluses, similar to the familiar approach of patient controlled analgesia, in addition to the continuous intrathecal infusion at a steady rate. Choice of Intrathecal Medication The only two pain- relieving medications that are approved for use in intrathecal pumps are morphine sulfate and ziconotide. Although morphine sulfate is a well- known medication that is familiar to both physi cians and patients, ziconotide is a unique calcium channel Figure 8.4 Stereotactic radiosurgery planning image showing position of the isocenter over the cisternal segment of the trigeminal nerve root.

morphine sulfate and ziconotide. Although morphine sulfate is a well- known medication that is familiar to both physi cians and patients, ziconotide is a unique calcium channel Figure 8.4 Stereotactic radiosurgery planning image showing position of the isocenter over the cisternal segment of the trigeminal nerve root. PAIN • 81 antagonist that is a synthetic form of the conotoxin pep tide, an analgesic substance derived from a cone snail, Conus magus. The side- effect profile of ziconotide is very different from that of opioids, and the best way to avoid development of psychiatric symptoms is to start with a low dose of medication and increase it slowly until desired pain relief occurs. There are many other medications that are suggested for intrathecal use despite lack of regulatory approval. These off- label medications include hydromorphone, bupiva caine, clonidine, fentanyl, and others. The recommenda tions on choice of intrathecal medications are updated every several years with published guidelines that summarize consensus of an expert panel on polyanalgesic use. The most recent guidelines list three first- line choices for neuropathic pain (morphine, ziconotide, and morphine + bupivacaine) and four first- line choices for nociceptive pain (morphine, hydromorphone, ziconotide, and fentanyl). Other medications and their combinations are listed as second through fifth lines in step- like algorithms. So, how does one decide whether the patient is a can didate for intrathecal drug delivery? It is considered a standard operating procedure to perform a trial of intra thecal morphine before implanting the intrathecal pump. Although many different approaches for trialing exist (sin gle bolus, multiple boluses, implanted catheter for either continuous infusion or periodic administration of escalat ing dose, epidural or intrathecal, medication only or placebo controlled), none is superior to the others. The simplest approach is to give a single intrathecal injection of a test dose of medication (usually, 1 mg of preservative- free morphine sulfate) through a regular lumbar puncture and then monitor the patient for 6 to 8 hours for degree of pain relief and for development of any side effects. Usually, more than 50% reduction in the pain intensity is considered a posi tive result that would justify proceeding with implantation of the pump. Less than 50% reduction in pain and no side effects may be an indication for repeating the trial with a higher dose of morphine, and good pain relief but development of side effects may be a reason to repeat the trial with a lower dose of medication. Implantation of the Pump Surgery for implantation of the pump is done under gen eral anesthesia— the patient is placed in a lateral position, and the incisions are planned over the midlumbar area to insert the catheter and over the abdominal wall to place the pump. Choice of the side depends on the presence of colostomy, nephrostomy, or surgical scars. Although in principle there are many different kinds of pumps, the usual choice is to use programmable pumps that allow one to change the daily dose of medication by simple reprogramming rather than by replacing the medication with a different drug concentration. T o reduce the risk for catheter fracture or migration, it is recommended to use oblique paramedian approach. It is also recommended to use fluoroscopic guidance and to enter the thecal sac below the level of L2— to reduce the risk for inadvertent injury of the conus. Special kinkresistant catheters are loaded with a flexible guidewire; they are inserted in cephalad direction through a T uohy- type needle. The catheter is usually advanced to a midthoracic level, and its location is controlled with fluoroscopy.

the level of L2— to reduce the risk for inadvertent injury of the conus. Special kinkresistant catheters are loaded with a flexible guidewire; they are inserted in cephalad direction through a T uohy- type needle. The catheter is usually advanced to a midthoracic level, and its location is controlled with fluoroscopy. After the guidewire is removed, the spontaneous flow of CSF from the catheter indicates patency of the tubing and sub arachnoid position of the catheter tip. It is important to anchor the catheter to the fascia— most often, an injectable anchor from the catheter kit is deployed over the catheter and sutured to the fascia with nonabsorbable suture. After that, the catheter is advanced to the pump pocket in the abdominal wall through a catheter passer. A standard shunt passer may be used for this purpose. The pump is filled with medication before it is attached to the catheter. After the catheter and pump are con nected, the pump is placed in its abdominal wall pocket, the catheter is coiled under the pump, and then the pump is secured to the underlying abdominal fascia with nonab sorbable sutures. Care is taken to avoid making the pump pocket too large because the pump may flip or migrate, or too small in order to avoid excessive tension of the tis sues. After the closure, the pump is programmed to start continuous infusion of the intrathecal medication, and an initial bolus is administered to clear the internal pump tubing and the entire catheter and then to fill them with the medication. Complications Despite being a straightforward surgical procedure, implantation of an intrathecal drug delivery system is associated with all kinds of complications, including surgical issues (e.g., hematoma; cord, or nerve root injury from needle insertion; CSF leak; seroma formation; wound infec tion), hardware- related issues (e.g., disconnection, catheter migration, catheter fracture [Figure 8.5], pump stall) and therapy- related issues (e.g., overdose resulting in respiratory depression, underdose resulting in withdrawal symptoms, itching and leg swelling from morphine administration, psychosis from ziconotide).

, hardware- related issues (e.g., disconnection, catheter migration, catheter fracture [Figure 8.5], pump stall) and therapy- related issues (e.g., overdose resulting in respiratory depression, underdose resulting in withdrawal symptoms, itching and leg swelling from morphine administration, psychosis from ziconotide). 82 • G OODMAN ’S N EUROSURGERY O RAL B OARD R E v IEW Intrathecal Catheter- Tip Inflammatory Mass (Granuloma) A separate concern is a possibility of development of an intrathecal catheter- tip granuloma (Figure 8.6). This is an inflammatory mass that is not associated with any kind of infection; it forms inside the intrathecal space interfering with medication release and distribution and becomes attached to the spinal cord and nerve roots, sometimes to a point of causing neurological deficits due to cord compression. Most catheter- tip granulomas occur with high concen trations of morphine sulfate, with use of morphine citrate, or with high doses of medications, but they are also described with other medications, including hydromorphone and fentanyl. Most granulomas remain asymptomatic and present as an incidental finding on follow- up imaging, but in some patients they manifest with gradual loss of efficacy (because the drug stays inside the granuloma and gets absorbed into the bloodstream instead of circulating in CSF) or neuro logical deficits due to direct compression of underlying spinal cord. The pumps that are being implanted today are considered MRI conditional and should not prevent the patient from undergoing MRI if clinically warranted. It is recommended to interrogate the pump after MRI to make sure the motor restarted after exposure to the strong mag netic field; some pump models require medication with drawal and reservoir refill after MRI. Asymptomatic granulomas do not require surgical intervention— if desired, the medication dose, concentra tion, infusion rate, or medication itself may be changed to stop granuloma progression. Minimally symptomatic granulomas would usually require stopping the pump and not restarting it until granuloma disappears, but if the patient is overtly symptomatic and presents with cord compression, it is recommended to decompress the cord with laminec tomy and duraplasty. Aggressive removal of granuloma, which was advocated in the past owing to concerns about a possible infectious nature of the inflammatory mass, is no longer recommended— partly because it is now confirmed that granulomas are sterile and partly because inflammatory reaction around the mass results in significant risk for neurological injury from granuloma resection. Figure 8.5 Sagittal reconstruction of the computed tomographic scan obtained during a “pumpogram” for suspected fracture and disconnection of the intrathecal catheter indicates contrast extravasation in an extraspinal location. The catheter enters the intrathecal space, but the medication does not reach the spinal canal because of a leak at the anchoring site— presumably due to needle penetration of the catheter at the time of its original implantation. Figure 8.6 Coronal reconstruction of a computed tomographic scan showing a hyperdense intrathecal lesion that formed around the tip of the intrathecal catheter. Such an intrathecal inflammatory mass may present with nerve compression symptoms or be an incidental finding during a routine imaging study.

nal implantation. Figure 8.6 Coronal reconstruction of a computed tomographic scan showing a hyperdense intrathecal lesion that formed around the tip of the intrathecal catheter. Such an intrathecal inflammatory mass may present with nerve compression symptoms or be an incidental finding during a routine imaging study. PAIN • 83 Overdose and Underdose Because the pump has to be refilled on a regular basis, there is a possibility of human error with missing the pump port (so- called pocket refill) or injecting the drug into a wrong port, or hitting the catheter with the needle and acciden tally injecting 3 to 6 months’ worth of medication directly into the spinal fluid. There may also be an error in using the wrong concentration of the drug or with programming a wrong infusion rate. The end result of this complication will be either a dramatic overdose of the patient with all signs of opioid toxicity (e.g., respiratory depression, somnolence, or coma) or development of withdrawal symptoms if the lower drug concentration was used, the pump was improperly programmed, or the catheter became obstructed or disconnected. T reatment of overdose includes initiating supportive measures such as intubation and respiratory support in the intensive care setting, administering proper medications (e.g., naloxone), stopping the pump, and emptying the pump reservoir. Withdrawal symptoms, if recognized on time, are best treated with administration of opioids orally or parenterally. Incomplete pain relief may indicate catheter obstruction; the other possible explanations would include development of tolerance and disease progression. Cordotomy But what if the patient’s life expectancy is shorter than 6  months? Implantation of an intrathecal pump in such a situation would not be appropriate for logistical and cost- effectiveness reasons. Instead, one would consider using a destructive option, which in this case of unilateral treatment- resistant pain from known malignancy would be a contralateral cordotomy. The target for intervention is the lateral spinothalamic tract that transmits pain infor mation from the contralateral side of the body toward the brain. This tract is located in the anterior (ventral) half of the lateral columns and projects anteriorly to the dentate ligament. Fibers of spinothalamic tract are somatotopically organized, and the cervical fibers project more medially within the tract than the thoracic, lumbar, and sacral fibers. Classical cordotomy consisted of a mechanical transec tion of the anterolateral quadrant of the spinal cord with a dedicated right- angle instrument (the cordotome). Most often, open cordotomy was being done at the upper thoracic level, and the surgical approach would include hemilami nectomy on the side of surgery contralateral to the side of pain. A less invasive alternative is CT- guided cordotomy in which the RF electrode is inserted through a special needle at the C1- C2 level and then a lesion is made in the location chosen based on the patient’s feedback in response to electrical stimulation. CT- guided cordotomy may be done on an outpatient basis; the entire procedure takes place in the CT scanner suite. For a CT- guided approach, we start the procedure with an injection of myelographic dye to define the borders of the spinal cord and to clearly visualize the cord on axial CT images. The infraauricular area is sterilely prepared, the needle insertion point is infiltrated with local anesthetic, and then the lateral upper cervical puncture is performed using standard landmarks. With the patient lying supine, the entry point is chosen 1 cm below the tip of mastoid process, and the needle is aimed slightly caudal and par allel to the floor.

epared, the needle insertion point is infiltrated with local anesthetic, and then the lateral upper cervical puncture is performed using standard landmarks. With the patient lying supine, the entry point is chosen 1 cm below the tip of mastoid process, and the needle is aimed slightly caudal and par allel to the floor. If needed, a series of images is obtained during the insertion process to confirm the needle direc tion. The needle used for cordotomy is a standard cut- tip needle, large enough to accommodate the special RF elec trode (Kanpolat CT cordotomy kit— KCTE); curved and closed- tip needles that are usually used for lumbar punctures (T uohy- and Whitaker- type needles) would not be appro priate. After the intrathecal space is entered, the stylet of the needle is replaced with the electrode, and the spinal cord entry is confirmed by the change in electrical impedance. The electrode is aimed anterior to the dentate ligament; its position is checked with axial CT images (Figure 8.7). T o confirm the correct position of the electrode tip, the electrical stimulation is used to elicit paresthesias. The curved tip of the electrode allows one to insert it in differ ent directions in order to reach the correct location inside the spinothalamic tract. For the patient described here, we would target the left (contralateral to the side of pain) lat eral spinothalamic tract and look for paresthesia coverage in the hip, pelvis, and right leg. After the paresthesias cover the painful region, a thermal RF lesion is created, usually with 70° to 72 ° C for 60 seconds. As a result of the lesion, the patient develops partial numbness in the area of the pain as well as significant and immediate improvement of pain. After this is achieved, the electrode and the needle are withdrawn, and the insertion point is covered with a bandage. Complications In general, unilateral cordotomy is a very safe interven tion. Development of neurological deficits is rare and may be detected early because the patient stays awake during the procedure and is asked to move the ipsilateral extremi ties every few minutes during each step of the procedure. Sensory deficits are usually well tolerated; motor deficits occur rarely and are unlikely to be severe or permanent. Theoretically, it is possible to injure vertebral artery or the upper cervical nerve root during the approach, but very

e the ipsilateral extremi ties every few minutes during each step of the procedure. Sensory deficits are usually well tolerated; motor deficits occur rarely and are unlikely to be severe or permanent. Theoretically, it is possible to injure vertebral artery or the upper cervical nerve root during the approach, but very 84 • G OODMAN ’S N EUROSURGERY O RAL B OARD R E v IEW small caliber of the needle and CT guidance are intended to minimize these risks. A serious problem may arise from a loss of ability to breathe when asleep (so- called Ondine’s curse), which is a feared complication of bilateral high cervical cordotomy. This is the main reason the cervical cordotomy is reserved for unilateral pain; bilateral pain is better treated with midline myelotomy, and if the patient with unilateral pain later develops pain on the opposite side (ipsilateral to cordotomy), then the second cordotomy may be done on the upper thoracic level so that the risk for central hypoventilation is minimized. CASE 3 CLINICAL PRESENT A TION A 27- year- old man presents with severe constant pain in his right arm. The pain started a few weeks after a major injury when the patient was thrown from his motorcycle after hitting a stationary object. At the time of injury, the patient was wearing a helmet. He had a short period of loss of consciousness and was unable to move his right arm. He was diagnosed with brachial plexus injury and despite thorough rehabilitation has not recovered any movements or sensation in the affected arm. Currently, however, the patient’s main concern is severe burning pain that involves his entire paralyzed arm but is worst from midshoulder all the way to his wrist and fingers. Neurological examination shows complete flaccid paralysis of the entire right arm with loss of all types of sensation from top of the right shoulder to the fingers as well as the armpit region consistent with C5- T1 dermatomes; there is a mild pupillary asymmetry suggestive of incomplete right Horner’s syndrome. DIAGNOSTIC WORKUP Although the pattern of pain clearly points toward involvement of the brachial plexus, one has to determine whether there is an actual avulsion of the nerve roots from the cord because the treatment of pain due to brachial plexus injury differs from that of brachial plexus avulsion. In general, the history and examination may be sufficient to make this differentiation with a high degree of certainty. T raction injuries, such as result from motorcycle crashes, and those with complete loss of both motor and sensory function tend to be associated with plexus avulsion (preganglionic lesion), whereas those due to direct injury of the supraclavicular region or the neck and with partial loss of sensation and movements would be associated with the plexus injury— including complete disruption of the plexus outside of the cervical spine (postganglionic lesion). Imaging may help in differentiating these conditions— presence of cervical and upper thoracic pseudomeningo cele at the foraminal level strongly suggests avulsion of the nerve roots. Such pseudomeningocele may be apparent on cervical spine MRI (Figure 8.8). An even more sensitive test would be cervical myelography or CT myelography, which may show subtle asymmetry of the neural root sleeves with dilation of these sleeves on the side of avulsion. Another way to confirm complete avulsion of the nerve roots from the spinal cord would be the electrophysiologic testing. Among many other signs that are well described in the literature, existence of normal sensory nerve action potentials in the setting of complete sensory loss indi cates preganglionic location of the injury and is essentially pathognomonic of brachial plexus avulsion.

spinal cord would be the electrophysiologic testing. Among many other signs that are well described in the literature, existence of normal sensory nerve action potentials in the setting of complete sensory loss indi cates preganglionic location of the injury and is essentially pathognomonic of brachial plexus avulsion. TREA TMENT OPTIONS The treatment of brachial plexus avulsion and nonavul sion injuries is divided into efforts to restore neurological function (and these are outside of this chapter scope) and management of chronic pain. As with other chronic pain conditions, the treatment starts with medications; and with exclusively neuropathic nature of the pain, the preference is given to anticonvulsants and antidepressants rather Figure 8.7 Computed tomography– guided cordotomy procedure: the myelographic contrast facilitates visualization of the spinal cord at the upper cervical level with a needle reaching the surface of the spinal cord and the electrode entering the ventrolateral quadrant of the spinal cord.

lsants and antidepressants rather Figure 8.7 Computed tomography– guided cordotomy procedure: the myelographic contrast facilitates visualization of the spinal cord at the upper cervical level with a needle reaching the surface of the spinal cord and the electrode entering the ventrolateral quadrant of the spinal cord. PAIN • 85 than antiinflammatory medications or opioids. Referral for interventional or neurosurgical treatment of pain is com monly initiated after the medical treatment fails. Although there are many reports of using spinal or peripheral neuromodulation for pain after brachial plexus avulsion, it turns out that all or most of them were in fact used for patients with postganglionic injuries. Currently, the only options suggested for postbrachial plexus avulsion pain are either destructive intervention through dorsal root entry zone (DREZ) myelotomy or supraspinal electrical stimulation, such as motor cortex stimulation or the deep brain stimulation of the periaqueductal or periventricular gray matter. Of these, the DREZ procedure has been used much more often and with much better overall success and therefore is considered a preferred treatment modality. DREZ Myelotomy The DREZ procedure involves destruction of the ipsilat eral dorsal horn at the level of avulsion. The rationale for the procedure is to silence those interneurons that became hyperactive as a result of the loss of afferent input, the condition that is called deafferentation and is thought to be the underlying mechanism of pain due to the brachial plexus avulsion. The substrate of the intervention is the most superficial layers of the dorsal horn of the spinal cord— Lissauer’s tract, substantia gelatinosa, and nucleus proprius of the dorsal horn (Rexed laminae I- V) have to be destroyed to obtain the pain- relieving effect. The surgery is done through either complete laminec tomy or, more frequently, hemilaminectomy at the side of pain. The patient is positioned prone, and the neck is kept in a neutral position, with the head supported by a three- pronged head holder. Unilateral exposure is suffi cient to identify the DREZ and perform the lesion to the desired depth. It is recommended to use intraoperative neurophysi ologic monitoring, mainly the somatosensory evoked potentials because they provide timely information about functional status of posterior columns and give the sur geon additional confidence regarding the directionality of destruction— posteromedial deviation from DREZ would increase the risk for postoperative posterior column dysfunction. The rostrocaudal extent of myelotomy is determined by the dermatomal distribution of pain. Most often, however, it is recommended to perform destruction from the last normal dorsal nerve rootlet above the level of the avulsion to the first normal rootlet below the level of the avulsion. So, if the entire brachial plexus is avulsed, as it appears to be the case in this patient, the lesioning should start at the lowest rootlet of C4 and continue to the most cephalad rootlet of T2 (assuming the avulsion occurred at levels C5- T1). Lesioning may be done using various surgical means; the standard approach uses an RF probe with an uninsu lated tip (Nashold thermocoagulation electrode) that is inserted into DREZ every 2 to 3  mm for thermal lesion ing. Other techniques use microsurgical transection with a scalpel blade, focused ultrasound delivered with a handheld probe, or CO2 laser. W e routinely use sharp bipolar forceps with coagulation of the spinal cord to the depth of 3 mm or so. Using this approach, the lesioning is performed under direct vision— use of the surgical microscope improves accuracy and provides better visualization and illumination of the procedure.

handheld probe, or CO2 laser. W e routinely use sharp bipolar forceps with coagulation of the spinal cord to the depth of 3 mm or so. Using this approach, the lesioning is performed under direct vision— use of the surgical microscope improves accuracy and provides better visualization and illumination of the procedure. Location and direction of the lesioning is chosen based on the individual spinal cord anatomy. Sometimes, a groove may be seen over the avulsed DREZ, but because the groove is not always present, we follow an imaginary line connect ing the normal dorsal rootlets above and below the level of avulsion. Direction of the approach aims toward the cen ter of the cord— it follows the position of the dorsal horn within the spinal cord (Figure 8.9). It is not uncommon to Figure 8.8 Axial T2- weighted magnetic resonance image of the cervical spine in a patient with traumatic right brachial plexus avulsion showing pseudomeningocele at the cervicothoracic level. In traction injury, such a finding may be considered pathognomonic for nerve root avulsion.

(Figure 8.9). It is not uncommon to Figure 8.8 Axial T2- weighted magnetic resonance image of the cervical spine in a patient with traumatic right brachial plexus avulsion showing pseudomeningocele at the cervicothoracic level. In traction injury, such a finding may be considered pathognomonic for nerve root avulsion. 86 • G OODMAN ’S N EUROSURGERY O RAL B OARD R E v IEW see grayish discoloration of the cord in the depth of lesion ing; after all, the procedure aims at the gray matter of the spinal cord. When the lesioning is completed, the dura is closed in standard fashion, and so are the soft tissues and skin. After awakening from anesthesia and extubation, the patient is kept in the hospital for 1 to 3 days, mainly to observe for any postoperative complications and to control postopera tive pain. The pain of laminectomy usually dominates the postoperative period; the initial pain disappears within a few days after the myelotomy. Complications Although generally considered a very safe procedure, DREZ myelotomy carries significant risk for postoperative neurological deficits. Almost all of them, however, are minor and do not cause major concerns for the patient because the improvement in pain is quite profound. Most published series mention a low incidence of postoperative neurological disturbances (<20%), such as sensory deficits related to the posterior column dysfunction (vibration and joint position sense), but in reality, most patients develop transient dyscoordination and difficulty with ambulation immediately after the surgery. Recovery from this may take up to 3 months; physical therapy and rehabilitation appear to speed up the recovery process. Other issues, such as weakness or bowel and bladder dysfunction, are quite rare. Widening of the area of numbness is not uncommon after DREZ myelot omy, but in most cases, the extent of sensory deficit returns to the preoperative baseline within 1 to 2 weeks. Other surgical complications, such as infection and CSF leak, are rare. CONTINGENCY PLAN What can be done if the pain returns or did not go away completely? The location of residual or recurrent pain would dictate the proper course of action. If the pain returns in the distribution of the most cephalad or most caudal dermatome (top of the shoulder, fingertips, or axilla), the DREZ lesion may be extended in the cephalad or caudal direction, respectively. If the pain never went away or returns in the entire arm, some more central interven tions may be considered, such as intrathecal drug infusion or electrical neurostimulation. The current algorithm for intrathecal management of neuropathic pain includes use of morphine, ziconotide, or morphine plus bupivacaine as first- line treatment. Electrical neuromodulation for a failed DREZ procedure includes stimulation of the spinal cord at a higher cervical level, motor cortex stimulation, or deep brain stimulation in the sensory thalamus or in the periaq ueductal or periventricular gray matter. CASE 4 CLINICAL PRESENT A TION A 52- year- old man presents with severe pain in his right foot for the last 20 months. The pain started almost imme diately after a trivial injury of the foot when a shopping cart rolled over it. At that time, the patient was diagnosed with a hairline fracture of his metatarsal bone and did not require any surgery. His foot was immobilized with a splint for few weeks, but despite apparent healing of the fracture, the patient continued to have pain. In addition to this, he developed severe sensitivity to any kind of touch, pressure, or manipulation of his foot, and after a year or so started exhibiting muscle atrophy, hair loss, toenail fragility, cold ness, and pale discoloration of the distal part of the right leg from midcalf down to his toes.

ued to have pain. In addition to this, he developed severe sensitivity to any kind of touch, pressure, or manipulation of his foot, and after a year or so started exhibiting muscle atrophy, hair loss, toenail fragility, cold ness, and pale discoloration of the distal part of the right leg from midcalf down to his toes. The pain improved only temporarily with various interventional treatments in the pain clinic. Currently, he is referred for neurosurgical evaluation by his pain specialist in hope of qualifying for some kind of pain- relieving surgery. DIAGNOSTIC IMPRESSION Making the diagnosis in such a case of peripheral pain may be challenging because the pain frequently does not fol low the clear anatomic border of a dermatome or sensory Figure 8.9 Postoperative axial magnetic resonance image of a patient with right brachial plexus avulsion who underwent right dorsal root entry zone (DREZ) myelotomy. The procedure is performed through a hemilaminectomy approach; the hyperintense lesion on T2- weighted image projects between the dorsal and lateral columns along the course of the DREZ and dorsal horn.

e image of a patient with right brachial plexus avulsion who underwent right dorsal root entry zone (DREZ) myelotomy. The procedure is performed through a hemilaminectomy approach; the hyperintense lesion on T2- weighted image projects between the dorsal and lateral columns along the course of the DREZ and dorsal horn. PAIN • 87 distribution of a specific peripheral nerve. In the past, the pain described here would be referred to as reflex sympa thetic dystrophy ; use of this descriptive term is related to frequently encountered sympathetic features (e.g., cold ness, discoloration, impaired perspiration) and develop ment of “dystrophy” or gradual atrophy of the soft tissues. Currently, this term is abandoned— instead, the cor rect diagnosis is complex regional pain syndrome (CRPS) type 1. The distinction between CRPS type 1 and type 2 is in involvement of a named nerve. Existence of such an involved nerve would be referred to as causalgia, or what is currently called CRPS type 2. Both types 1 and 2 may be sympathetically maintained or sympathetically independent; presence or absence of autonomic features does not change the diagnosis or treatment. There is no particular tool to diagnose CRPS. Structural imaging is expected to be normal; in advanced cases, the imaging would confirm the presence of tissue atrophy on the symptomatic side. Thermography was proposed as a diagnostic test for CRPS because most patients would have a pronounced asymmetry in tissue temperature between the normal and affected sides. T R E AT M E N T The best treatment for CRPS is mobilization of the affected extremity. CRPS symptoms do indeed resemble find ings observed in so- called disuse syndrome, as one would develop, for example, after prolonged wearing of a cast. However, it is difficult to participate in therapy while suffering from uncontrollable pain. Therefore, it is recommended to proceed with aggressive pain management before ther apy is initiated. The surgery referral would be appropri ate if the patient fails to improve in response to standard medical treatment with anticonvulsants and antidepres sants. Usually, the patient would receive local anesthetic or sympatholytic blocks from a pain specialist before being referred for surgical treatment. Current surgical treatment of pain in CRPS is straightforward. Although there is a destructive option of sympa thectomy, it is rarely, if ever, used now for anything other than hyperhidrosis. Over the past several decades, electri cal neuromodulation, particularly spinal cord stimulation (SCS), has become an accepted surgical treatment of CRPS pain, and a published randomized prospective study sup ported its efficacy. In fact, SCS is the most common surgi cal procedure that is done specifically for treatment of pain, and CRPS is one of its most established indications. Spinal Cord Stimulation There are several criteria that qualify a patient for a SCS procedure. These include presence of chronic (>6 months), severe (numerical rating scale of 6 or higher), disabling (affecting one’s functionality) pain; established medical diagnosis; favorable psychological evaluation results; the patient’s ability and willingness to operate the implanted device; preserved sensation in the painful region; and finally— and perhaps most important— successful trial of stimulation. Most prerequisites can be checked during ini tial evaluation; psychological evaluation requires separate referral; the patient education prepares him for the surgi cal interventions. Finally, the trial of SCS allows one to determine its efficacy and the severity of side effects, if any, related to the stimulation. In cases of documented CRPS, it may make sense to intervene before the usual 6 months of conventional treat ment are over.

patient education prepares him for the surgi cal interventions. Finally, the trial of SCS allows one to determine its efficacy and the severity of side effects, if any, related to the stimulation. In cases of documented CRPS, it may make sense to intervene before the usual 6 months of conventional treat ment are over. Early physical therapy is the key to success, but as mentioned earlier, it is hardly possible for patients to participate in therapy as long as they are in severe pain. The target for SCS is the dorsal column. Despite many discussions and arguments, the gate control theory of pain that postulates an ability to suppress the pain by deliver ing nonpainful information to the interneurons of the spinal cord is still the most likely explanation, at least for paresthesia- based SCS. The craniocaudal location of the SCS target is determined by the location of pain— those with pain in the lower back and legs would usually achieve best results with stimulation of the lower thoracic region (between T7 and T10), pain in feet and perineal area would require stimulation below T10, and the pain in arms and hands would respond to SCS at the mid and low cervical levels. There are many types of SCS systems— percutaneous and paddle- type leads (Figure 8.10), rechargeable and non- rechargeable devices, constant current– and constant voltage– based systems— but most of them use implant able pulse generators and not RF receivers, which have become obsolete over the years. Choice of device (per cutaneous vs. paddle) is determined by several practical nuances— patients with lack of true epidural space due to previous laminectomy, those with tight stenosis at the level of stimulation, and those with history of multiple device migrations would do better with a paddle- type electrode that is implanted through a small laminectomy or lami notomy. Other patients probably would do just as well with percutaneous electrodes that are much less invasive and may be advanced over multiple spinal levels through a simple epidural entry with a T uohy- type needle. However,

o better with a paddle- type electrode that is implanted through a small laminectomy or lami notomy. Other patients probably would do just as well with percutaneous electrodes that are much less invasive and may be advanced over multiple spinal levels through a simple epidural entry with a T uohy- type needle. However, 88 • G OODMAN ’S N EUROSURGERY O RAL B OARD R E v IEW relative unfamiliarity of neurosurgeons with percutaneous implantation technique may be another reason that most electrode leads implanted today by neurosurgeons are of the paddle type. Spinal Cord Stimulation Trial The usual approach to SCS is to do a trial with percutane ous electrodes (unless the patient has one of the issues listed earlier that requires paddle electrode), and then, during the 5 to 7 days of the trial, to determine the patient’s response. The electrodes are placed with the patient awake in prone position under fluoroscopic guidance, and the intraopera tive testing helps to determine paresthesia coverage with the goal of covering all painful areas with stimulation- induced paresthesias in a concordant fashion. If the electrodes are intended to be temporary, they are simply sutured or taped to the skin for the duration of the trial; these electrodes are removed at the end of the trial, and the permanent implantation includes placement of a new electrode that is then connected to the generator. Another option is to keep the same electrodes that were used for the trial— in this case, they should be anchored to the fascia and con nected to a set of extensions that is then tunneled under the skin to a distant exit site. At the end of this “tunneled trial, ” the incision is reopened, the extension cables are cut and discarded, and the electrodes are tunneled to the gen erator pocket. The main advantage of the tunneled trial is that the electrodes that were tested during the trial remain in use for long- term stimulation. The disadvantages of this approach are that the patient requires electrode placement in the operating room, as opposed to the doctor’s office or the procedure room, and that another trip to the operating room is required even if the trial fails because the anchored electrodes need to be taken out by reopening the original incision. Implantation of a Permanent Device This stage of surgery is usually done under general anesthesia in prone or lateral position depending on the chosen location for the generator. W e prefer placing the generator into the abdominal wall, where it is easier to program and charge— this requires lateral positioning of the patient; others use the flank or buttock for generator placement, and this may be done in prone position. The electrode is con nected to the generator after it is tunneled to the generator pocket. The pocket depth is dictated by the need for recharg ing: non- rechargeable devices may be placed up to 4 cm deep under the skin, but rechargeable devices should be no more than 1.5 to 2 cm under the skin surface. The size of the generator pocket should match the size of the generator. Making the pocket too large may result in device rota tion and possible seroma formation. Making the pocket too small would produce significant tension of the tissues and elicit pocket- site pain after the implantation. Most companies making devices also make plastic templates that may help in sizing the pocket before the generator insertion. A B C D Figure 8.10 Spinal cord stimulation electrode leads of different types. A: Percutaneous eight- contact cylindrical electrodes. B: Percutaneously insertable narrow- paddle electrodes. C: Three- column paddle- type electrode. D: Five- column paddle- type electrode.

izing the pocket before the generator insertion. A B C D Figure 8.10 Spinal cord stimulation electrode leads of different types. A: Percutaneous eight- contact cylindrical electrodes. B: Percutaneously insertable narrow- paddle electrodes. C: Three- column paddle- type electrode. D: Five- column paddle- type electrode. The bayonet- shaped insertion tool is intended to advance and steer the paddle- type electrode while avoiding putting direct pressure on the thecal sac.

izing the pocket before the generator insertion. A B C D Figure 8.10 Spinal cord stimulation electrode leads of different types. A: Percutaneous eight- contact cylindrical electrodes. B: Percutaneously insertable narrow- paddle electrodes. C: Three- column paddle- type electrode. D: Five- column paddle- type electrode. The bayonet- shaped insertion tool is intended to advance and steer the paddle- type electrode while avoiding putting direct pressure on the thecal sac. PAIN • 89 Postoperative Care Current SCS devices allow for multiple programs to be used based on the patient’s preference. These programs include different combinations of stimulating contacts, different frequencies and pulse widths of stimulation, and different amplitude ranges within which the patient can vary the intensity of stimulation. Nevertheless, it is common for SCS patients to undergo reprogramming sessions in the postoperative period, usually once every 1 to 3 weeks, to optimize the coverage and achieve reliable pain relief. Currently, there are at least a half dozen SCS devices that may be used for most indications. Most of them have unique features, such as adaptable stimulation that changes based on the patient’s position, ability to deliver burst or very high- frequency stimulation, special programming approaches that allow for independent current delivery to multiple contacts, different paddle configurations with two to five columns, and smaller paddle electrodes that may be implanted through percutaneous approach. Most devices implanted today have limitations in terms of MRI compatibility. None is MRI safe, but some have conditional approval for use in 1.5T or 3T scanners for certain parts of the body or even for whole- body scanning as long as certain conditions are met. However, at the time of this writing, there are only two MRI- compatible paddle electrodes available. The actual treatment of CRPS starts after SCS implantation. The pain relief allows the patient to start intensive physical therapy that may reverse trophic changes and muscle atrophy associated with this condition. Eventually, if the pain subsides, the stimulation may be stopped. Subsequently, the device may be safely removed if it has not been used for 6 months or longer. Complications Use of SCS devices is associated with several types of com plications and side effects. Some of them are procedurerelated and include hematoma, infection, damage to the nervous tissue (spinal cord or nerve roots), and internal or external CSF leak that may occur if the underlying dura was inadvertently violated during electrode insertion. Others may be related to the hardware and surgical technique; these include device migration, fracture, and disconnection. Finally, there is a group of therapy- related effects, includ ing loss of stimulation benefits, development of unpleasant paresthesias, pain from stimulation, and discomfort during recharging. Most of these complications are easy to recog nize and treat. Loss of stimulation may be remedied by reprogramming; this may also be used to eliminate unpleasant paresthesias or stimulation- induced discom fort. Sometimes, changing electrode polarity will suffice, but in other cases, the entire stimulation paradigms have to be changed to either higher frequency, bursting mode, or wider pulse widths. Pocket hematoma or seroma may have to be drained. Migrated electrodes have to be revised and repositioned; repeated migration of percutaneous leads may be a reason to switch to a paddle- type electrode, but one has to keep in mind that paddle leads are not immune to migration either. Infection should be treated cautiously, and superficial infection of soft tissues may be resolved with a short course of antibiotics. Deep infections and those involving the back incision usually indicate that the hardware is involved.

e, but one has to keep in mind that paddle leads are not immune to migration either. Infection should be treated cautiously, and superficial infection of soft tissues may be resolved with a short course of antibiotics. Deep infections and those involving the back incision usually indicate that the hardware is involved. In situations like this, it is recommended to remove the entire system and replace it with a new one a few months later after the infection has cleared. Hematoma is of particular concern when it occurs in the epidural space. Development of cord compression symp toms after electrode implantation should prompt urgent surgical decompression (Figure 8.11). Sometimes, the symptoms are caused by the paddle, which behaves as a mass lesion and therefore has to be removed during the decom pressive intervention. However, it is also possible that the cord was damaged during the insertion procedure, and the deficits may be a result of the cord edema. T o minimize this risk, it is recommended to use intraoperative neurophysi ologic monitoring during asleep implantation cases. It is also advisable to avoid aggressive epidural manipulation, to Figure 8.11 Postoperative axial thoracic computed tomography scan showing an epidural hematoma that is pushing the electrode against the thecal sac in a patient with newly developed neurological deficit several hours after uneventful insertion of the spinal cord stimulation electrode.

gressive epidural manipulation, to Figure 8.11 Postoperative axial thoracic computed tomography scan showing an epidural hematoma that is pushing the electrode against the thecal sac in a patient with newly developed neurological deficit several hours after uneventful insertion of the spinal cord stimulation electrode. 90 • G OODMAN ’S N EUROSURGERY O RAL B OARD R E v IEW start with adequate exposure in longitudinal and transverse directions, and to obtain preoperative imaging of the operated area (in this case— thoracic spine) to avoid any sur prises such as herniated disks or stenosis. CONTINGENCY PLANS If conventional SCS does not provide expected pain relief, one may explore alternative options. In addition to new stimulation paradigms (e.g., high frequency, high density, burst stimulation), there is an option of intrathecal drug delivery, similar to what was discussed earlier. There is anecdotal evidence that addition of local anesthetic, such as bupivacaine, to the opioid medication may be more effec tive in pain control in CRPS cases. Anecdotal reports have been made of destructive sur gery and central neuromodulation procedures, such as deep brain stimulation, used for CRPS treatment. The results, however, have been somewhat discouraging, particularly in advanced and more chronic cases.