Indications	Probability of Success
Postlaminectomy syndrome/failed back surgery syndrome (FBSS)	50%-60% with >50% pain relief
Complex regional pain syndrome (CRPS)	67%-84%
Chronic critical limb ischemia (CCLI) and pain	70%-80% with >75% pain relief
Angina	No data available
Abdominal/visceral pain syndromes
Brachial plexitis/neurogenic thoracic outlet syndrome
Phantom limb pain
Intractable pain secondary to spinal cord injury
Mediastinal pain
Cervical neuritis
Postherpetic neuralgia

Anatomy

The spinal canal contains several neural and nonneural structures that can be stimulated electrically. The properties of the intraspinal contents are similar to an inhomogeneous conductor. Knowledge of the different type of responses and their correlation with the underlying anatomical substrate is extremely important in implementing strategies for SCS.

Stimulation of large myelinated afferent fibers may occur in four different areas: the dorsal root, the dorsal root entry zone (DREZ), the dorsal horn, or the dorsal columns (Fig. 5-1). Electrical stimulation of these structures results in an ipsilateral tingling paresthesia, but changes to discomfort and pain with increasing stimulation voltage. Particularly, activation of the dorsal roots causes a radicular paresthesia at a particular dermatomal level, whereas stimulation of the dorsal columns causes paresthesia in areas of the body caudal to the level of the electrode. In the cervical and thoracic cord the dorsal columns are usually about 3 to 5 mm wide, largest at the level of the cervical enlargement, and smallest in the midthoracic spine.

Fig. 5-1 Anatomy of the spinal canal. At T9 the lower thoracic fibers are located laterally within the dorsal columns. Some of these fibers are in close proximity to the dorsal root entry zones. Dorsal root fibers are larger, curved, and more easily activated than dorsal column fibers. Within the lower thoracic spine, the highly conductive dorsal cerebrospinal fluid (dCSF) layer is thickest where spinal cord stimulation leads are typically placed for back pain therapy. The dCSF can diffuse the stimulation field, increasing the likelihood that the dorsal roots will be inadvertently stimulated.

The single most important factor in determining stimulation parameters is the width of the cerebrospinal fluid (CSF) space dorsal to the spinal cord (Fig. 5-2). Increasing thickness of the dorsal CSF (dCSF) layer leads to early stimulation of the dorsal root fibers instead of the dorsal column fibers. This is largely because rootlet fibers, which are immersed in well-conducting CSF, have high conductivity at the DREZ. Therefore the stimulation threshold for the segmentary system ranges from 0.1 V to 0.5 V, whereas activation of the dorsal columns occurs at a threshold that is at least 0.5 V to 1 V higher.

Fig. 5-2 Relationship between spinal level and perception threshold/dorsal cerebrospinal fluid thickness.

Differentiating between stimulation of the dorsal root, DREZ, or dorsal horn can be exceedingly difficult. One can expect dorsal root stimulation with laterally placed electrodes; stimulation of the DREZ or dorsal horn is more likely if the electrode is placed near midline. In the latter group stimulation leads to segmentary paresthesias that are rapidly followed by activation of the dorsal columns with a small voltage increment. Interestingly, minute changes in mediolateral position of the electrode have been shown to lead to movement of the paresthesias in a W two-dimensional pattern ¹ as shown in Fig. 5-3.

Fig. 5-3 Migration of paresthesias with mediolateral adjustment of stimulation.

Modified from Oakley JC et al: Transverse tripolar spinal cord stimulation: results of an international multicenter study, Neuromodulation 9(3):192-203, 2006.

If electrical stimulation reaches the ventral cord, motor structures are affected, and consequently muscle contractions are seen. Much like dorsal cord stimulation, activation of the ventral root or motor neurons causes muscular contractions in a myotomal distribution, whereas activation of the descending corticospinal pathways results in diffuse muscle contractions caudal to the level of the electrode. In the vast majority of patients stimulation over the dorsal columns is not complicated by simultaneous activation of the motor system unless a significantly higher voltage is implemented. Nevertheless, there have been cases in which activation of the motor structures has been seen to occur at the same threshold as the sensory system; therefore selective stimulation of the dorsal column could not be successfully maintained.²

Understanding the somatotopy of the spinal cord is paramount to knowing the technical aspects of implantation. A basic tenet of SCS is to create an area of paresthesia that overlaps the patient’s region of pain.³ Therefore correlation of the somatotopy and the level of the spinal cord is necessary. Barolat and colleagues⁴ have published extensively on the mapping of the spinal structures, which led to a database correlating areas of sensory response to levels dorsal SCS.⁴

High cervical regions such as C2 can cover regions of the posterior occiput and occasionally the lower jaw. C2-C4 stimulation provides coverage of the neck, shoulder, and upper extremities, as well as the face as a result of involvement of the descending nucleus of the trigeminal nerve. As one moves to the lower cervical region such as C5-C6, stimulation affects the entire hand. To cover the anterior chest wall or the axilla, an electrode placed toward C7-T1 is necessary.

More commonly an implanter seeks coverage of the lower extremities. Lateral placement at T11-T12 covers the anterior thigh, whereas placement at T11-L1 can cover the posterior thigh. Although coverage of the foot as a whole can be achieved by stimulating these same areas, it is difficult to cover the sole of the foot, which may require insertion on the lumbar L5 or S1 nerve roots. Low back pain is very difficult to cover because midthoracic stimulation can also affect the chest and abdominal wall. In general, midline placement of the electrode at T8-T10 seems to provide the best outcomes.

Recently this paradigm has been challenged by case series demonstrating successful four-limb paresthesia in patients receiving only cervical stimulation.⁵ In general, this effect was achieved with stimulation with longer pulse widths and lower frequencies. Although the mechanism is not yet well understood, some have hypothesized that thinner dorsal CSF thickness in the cervical spine may allow for the stimulation of deeper medial and midline fibers that correspond to the lower extremities.

Basic Science

The enthusiasm with SCS began with the introduction of the gate control theory for pain control by Melzack and Wall ⁶ in 1965. They noted that stimulation of large myelinated fibers of peripheral nerves resulted in paresthesia and blocked the activity in small nociceptive projections. In other words, appropriate stimulation of a competing afferent signal can effectively block an existing pain signal.

Although a large body of work has been published, the exact mechanisms of action of SCS remain unclear. The computer modeling work of Coburn,^7,⁸ Coburn and Sin,⁹ Holsheimer and associates,¹⁰ Holsheimer and Strujik,¹¹ and Holsheimer and Wesselink¹² have shed some light, at least theoretically, on the distribution of the electrical fields within the spinal structures. It is clear that stimulation on the dorsal aspect of the epidural space creates complex electrical fields that affect a large number of structures, including afferents within the peripheral nerve, dorsal columns, or supralemniscal pathways. Mechanisms at distinct levels of the central nervous system apart from the spinal cord are believed to contribute to the effects of SCS. Such mechanisms include antidromic action potentials that pass caudally in the dorsal columns to activate spinal segmental mechanisms in the dorsal horns and action potentials ascending in the dorsal columns activating cells in the brainstem, which in turn might drive descending inhibition. Recent work by Schlaier and associates¹³ demonstrated changes in cortical excitability with SCS, which is believed to be N-methyl-d-aspartate (NMDA)-related neuroplasticity at the supraspinal level. At the chemical level, animal studies suggest that SCS triggers the release of serotonin, substance P, and γ-aminobutyric acid (GABA) within the dorsal horn.¹⁴^–¹⁶

Indications/Contraindications

As mentioned previously, SCS has been used for a variety of pain conditions and is particularly indicated for pain of neuropathic origin rather than for nociceptive pain. It is important to realize that neurostimulation is a treatment option along the continuum of pain control.

Another important consideration is the need for careful psychological screening in the selection of candidates for SCS procedure.¹⁷^–²⁴ This becomes particularly important because mood disorders can alter pain reporting and perception by a patient. Patients who have frank psychiatric disorders, excessive depression, anger, or unrealistic expectations may be inappropriate for SCS; however, some of these patients may become favorable candidates for SCS after appropriate psychiatric therapy. The most common indications for SCS follow.

Postlaminectomy Syndrome/Failed Back Surgery Syndrome

Postlaminectomy syndrome is vaguely defined. The term has included pain localized to the center of the lower lumbar area, pain in the buttocks, persistent radicular pain, or diffuse lower extremity(s) pain. Arachnoiditis, epidural fibrosis, radiculitis, microinstability, recurrent disc herniations, and infections have been perpetrated in the etiology of this syndrome. Although SCS is accepted in the treatment of leg pain, its widespread use for relief of pain in the lower lumbar area still remains to be defined.

Complex Regional Pain Syndrome

CRPS is characterized by gradually worsening severe pain, swelling, and skin changes of the hands, feet, elbows, or knees. It may be secondary to injury (CRPS type II) or may not have a clear etiology (CRPS type I). Implementation of SCS in these patients may be more difficult than with any other patient group, partly because of the ill-defined nature of the pain. In addition, there is an increased possibility of aggravating the original pain or causing new pain or allodynia at the site of implanted hardware. Furthermore, the pain may also spread to other body parts, thus making lasting coverage extremely challenging.

Chronic Critical Limb Ischemia and Pain

In 1973 Cook and Weinstein ²⁵ were the first to suggest that the indications for SCS might extend beyond intractable pain control. They observed a group of patients with multiple sclerosis who underwent SCS to treat their chronic pain. Unexpectedly the patients experienced not only pain relief but also an improvement in mobility and sensory and bladder function. They incidentally found an apparent improvement in lower limb blood flow and subsequently used SCS in patients whose primary problem was peripheral vascular disease (PVD).²⁶ These patients were found to have relief of rest pain, increased skin temperature, improved plethysmographic blood flow, and healing of small cutaneous ulcers.

Angina

The role of SCS in the management of refractory angina pectoris seems to be very promising. There are well-documented reports in the literature revealing uniformly good results in the relief of angina pain.²⁹^–³³ Further, the results have been maintained in long-term follow-up and have been substantiated by a reduction in the intake of nitrates as well. Interestingly, other findings have supported the evidence that SCS has effects that go beyond pain relief. The observations that there is less ST segment depression and that the exercise capacity, the time to angina, and the recovery time all improve with stimulation may suggest that there is a reduction in ischemia. In a positron emission tomography study, a redistribution of myocardial flow in favor of ischemic parts of the myocardium has been demonstrated as a long-term effect of SCS, both at rest and after pharmacological stress induction.³⁴

Moreover, since the relation between pain and myocardial ischemia has not been fully clarified, we do not know whether the pain relief is caused by direct depression of nociceptive signals in the spinal cord or whether there is secondary gain from a reduction in the ischemia.^35,³⁶ On the one hand, a significant amount of work by Foreman¹⁴ has shown that dorsal column stimulation inhibits the activity of spinothalamic tract cells, which are typically activated by cardiac sympathetic afferents or intracardiac bradykinin. On the other hand, the effects of stimulation might be equivalent to those of a sympathectomy in that it inhibits a hyperactive sympathetic system. This latter mechanism has been shown experimentally in the rat by Linderoth and associates.³⁷

Brachial Plexitis/Neurogenic Thoracic Outlet Syndrome

SCS has also been described for pain secondary to either brachial plexitis or neurogenic thoracic outlet syndrome. Most of these patients complain of pain that affects the upper extremity, the shoulder, the trapezius, the axilla, and/or the anterior upper chest wall. Unfortunately literature is relatively lacking for this indication.

Other Indications

Early studies in the 1980s demonstrated benefits of SCS for urinary incontinence, with consequent increases in continence and substantial reduction in residual urine volumes.³⁸^–⁴² These effects were found with stimulation of ventral sacral roots, a method that has also been shown to improve bowel function in patients with spinal cord injury.⁴³ Trials in the 1970s reported initial benefits of this modality for phantom limb pain, but few of these patients experienced lasting results with long-term follow-up.⁴⁴ Scattered reports also exist demonstrating the benefit of SCS in the treatment of intractable pain secondary to spinal cord injury pain, mediastinal pain, cervical neuritis, and postherpetic neuralgia.

Equipment

Electrical stimulation consists of rectangular pulses delivered to the epidural space through an implanted electrode via a power source. Electrodes were initially all unipolar in nature, but subsequently bipolar and tripolar arrays were developed. The initial radiofrequency (RF)-driven passive receivers have given way to implantable pulse generators (IPGs), which were first introduced in the mid-1970s. In 1980 the first percutaneous quadripolar electrode was produced; it could be reprogrammed noninvasively through an external transmitter.⁴⁵ Subsequently IPGs that can be both transcutaneously charged and programmed have been developed. This most recent advance is now leading a renewed interest in the possibility of using special electrode arrays in the delivery of electrical stimulation to the spinal cord.

Percutaneous Electrodes

Percutaneous electrodes, otherwise known as wire electrodes, are particularly appealing because they can be inserted without much dissection and are more appropriate for performing a trial to assess candidacy for a permanent implant. During implantation these electrodes can be advanced over several segments in the epidural space, allowing testing of several spinal cord levels to assess for optimal electrode position. After the trial period these electrodes can easily be removed in the physician’s office. Contemporary percutaneous electrodes are slim (i.e., only a few millimeters in diameter) and contain four (quadripolar) or eight (octipolar) contacts with various spacings. These percutaneous electrodes also come in varying lengths, which differ by manufacturer. Last, extension cables are occasionally necessary to bridge the distance from the spinal entry point to the pocket in which the battery will reside.

Electrode choice generally depends on how many segments of the spinal cord are to be covered. Although electrodes with larger spacing allow for broader coverage, closer spacing permits better steering and electric field shaping. The general trend is to use one or two quadripolar electrodes for limb pain and one or two octipolar electrodes for axial pain. Even insertion of three electrodes is being explored for better steering of current.⁴⁶

A major disadvantage of percutaneous electrodes is their tendency to migrate—secondary to their inherent flexibility, which is necessary for their insertion through a Tuohy needle and their cylindrical shape. These factors can lead to migration even months after implantation. As a result, patients with percutaneous leads have reported a greater positional variance in their paresthesia. To address this issue, some percutaneous electrodes have been introduced that require a stiffening stylet for introduction. Another disadvantage is that percutaneous electrodes are less energy efficient than plate electrodes. The electrical current is distributed circumferentially around the electrode and consequently results in greater shunting of current.

Recently ANS (A St. Jude company, Plano, Tex) has introduced a slim-line plate type electrode that can be inserted percutaneously. The broader electrode base provides a surface in which fibrosis should lessen the risk of caudal electrode migration. The slimmer profile of the electrode might also have advantages in the cervical spine where spinal cord compression might be an issue. Finally, the design emulates that of a miniplate lead since the contacts are on one side with the other side being insulated, leading to a more energy efficient system (Fig. 5-4).

Fig. 5-4 Slim-line electrode.

Paddle Electrodes

Paddle or laminotomy electrodes are offered in many sizes, shapes, lengths, spacing, and configurations of electrodes. They are primarily offered as single- and dual-column configurations; with three (tripole)- and five-column models being introduced more recently (Figs. 5-5 and 5-6). With regard to their shape, these electrodes may come with curved or hinged leads to help facilitate insertion.

Fig. 5-5 Tripole and Pentad electrode paddles.

Fig. 5-6 Types of plate electrodes.

The main advantage of these electrodes resides in their more inherent stability in the dorsal epidural space and lesser propensity to migrate. Some preliminary data by North and associates ⁴⁷ also suggest a broader stimulation pattern and lower stimulation requirements. As alluded to previously, paddle electrodes are also energy efficient in delivering electrical stimulation.

There is some theoretical evidence that shaping of the electrical field is possible with more complex electrode arrays. Current available tripole electrodes and their dimensions are shown in Table 5-2. Holsheimer and colleagues⁴⁸ concluded that the transverse tripolar system enabled finer control of paresthesia because of its ability to more accurately shape the stimulating electromagnetic field. Specifically the transverse tripolar configurations allow for stimulation of the dorsal columns at higher amplitudes before the dorsal roots begin experiencing clinically significant stimulation. These electrodes are composed of a central cathode flanked by lateral anodes, which modulate the field medially or laterally, depending on their programmed voltage. In other words, the current delivered to the spinal cord can be steered not by moving the electrode but instead by simply modifying the voltage ratio between the central cathode and two flanking anodes.¹ The lateral anodes create a shielding effect such that the dorsal roots are protected from stimulation; deeper activation of the dorsal columns is achieved with higher voltages (Fig. 5-7). This feature ultimately results in a wider therapeutic range, wider paresthesia coverage, and a greater probability to fully cover the painful area with paresthesia. Possible programming configurations of these tripole electrodes are shown in Fig. 5-8.

Table 5-2 Tripole Electrodes

	MDT Specify 5-6-5	ANS Lamitrode Tripole 16C
Electrode length (mm)	4.0	Mid: 4.1 Out: 5.8
Electrode width (mm)	1.5	Mid: 2.6 Out: 1.8
Longitudinal ctr. to ctr. (mm)	9.0	Mid: 7.0 Out: 7.0
Lateral ctr. to ctr. (mm)	3.0	3.2
Electrode area (mm²)	6.0	Mid: 10.7 Out: 10.4

Fig. 5-7 Electromagnetic field comparison between traditional and tripole electrodes.

Fig. 5-8 Tripole programming configurations.

Although increasing the number of contacts allows for finer control of the electromagnetic field, there is also a significant increase in power consumption and an even greater increase in the complexity of programming. The number of possible configurations is n2ⁿ, where n is the number of electrodes. Therefore with two contacts the total number of configurations possible is eight; with four contacts, sixty-four; and with sixteen contacts, it increases exponentially to reach a number in the millions.

Implantable Pulse Generators

The totally implantable pulse generator contains its own lithium battery. Activation and modification of stimulation parameters can occur through an external transcutaneous telemetry device, which the patient can carry. More extensive programming of the system can be achieved through a small portable unit that can be managed by the physician. These generators allow stimulation with fine resolution increments of 0.05 V and with varying rates and pulse widths.

Some systems also use multiple channels, which allow for treatment of multiple, distinct areas of the body using a single IPG. Once programmed, the patient can also select a particular channel at any particular time to best cover his or her symptoms.

Battery life varies with usage and with the used parameters such as voltage, rate, and pulse width. These factors, particularly voltage, are affected by patient factors such as epidural scarring and accommodation to stimulation. With typical use these batteries last between 2.5 to 4.5 years. Replacement of the battery requires a surgical procedure, which is usually performed on an outpatient basis.

IPG selection is based on many variables. From a practical standpoint, the first and foremost reason might be the size of the patient. Although larger batteries have a longer life, they obviously have more bulk and can often become the source of significant patient complaints. Although the IPG can be implanted into the buttock, abdomen, or subclavicular area, implantation in the buttock has multiple advantages. This location makes it easier to tunnel the electrode leads to the battery; with a patient placed prone, repositioning is not required to reach the buttock region.

Rechargeable Implantable Pulse Generators

Rechargeable systems have now become available. Medtronic’s device, known as the Restore Rechargeable Neurostimulation System, uses a battery with an estimated 9-year total life span and takes about 6 hours to fully recharge. Advanced Neuromodulation Systems’ Eon device has a battery life that is currently estimated to be 7 years. Boston Scientific’s Precision device has a battery life estimated to be 5 years. A detailed comparison of the features of these rechargeable batteries is shown in Table 5-3.

Table 5-3 Rechargeable Batteries

In an analysis of cost between rechargeable and nonrechargeable systems, Hornberger and associates ⁴⁹ found that the former would require 2.6 to 4.2 fewer replacement procedures, resulting in a total lifetime savings ranging from $104,000 to $168,833. With regard to overall efficacy, a 12-month study looking specifically at the Restore rechargeable IPG demonstrated that the system was easy to use and that all patients were able to recharge it successfully. Moreover, both patient and physician satisfaction was high, with significant improvements in pain, quality of life, and functional status.⁵⁰

Technique

In 1967 Shealy and Cady ⁵¹ and Shealy, Mortimer, and Reswick⁵² inserted the first dorsal column stimulator in a human suffering from terminal metastatic cancer. Subsequently electrodes have been implanted using a variety of techniques: via a laminectomy in the subarachnoid space⁵²; between the two layers of the dura, or in the epidural space, either dorsal or ventral to the spinal cord.⁵³^–⁵⁵ Subsequently less invasive percutaneous techniques were introduced.⁵⁶ Current practice indicates the placement of electrodes in the dorsal epidural space and connecting them to an implantable pulse generator usually placed in the buttock. However, before a permanent system is implanted, patients typically undergo a trial period of 3 to 10 days, during which time stimulation is provided by percutaneous electrodes to determine their overall efficacy.⁵⁷

Placement of Percutaneous Electrodes

To place percutaneous electrodes, the patient is positioned in a comfortable prone position on a fluoroscopy table with a pillow underneath the abdomen. This position creates some kyphosis, which may facilitate electrode insertion.

A fundamental consideration in choosing the level of entry is that several centimeters of the lead have to lie in the epidural space to ensure maximal stability of the electrode and minimize unwanted migration. To ensure this, insertion must take place at least two spine segments below the desired target. For cervical placement one must bear in mind the cervical cord enlargement and insert the electrode below the T1-T2 level when possible. Insertion is performed under fluoroscopy; this equipment must be ready to function in both the anteroposterior and lateral planes at the time of needle insertion. A Tuohy needle is inserted with as shallow an angle as possible to minimize the risk of electrode breakage and a paramedian approach to avoid friction with the spinous process.

Although tactile feedback is important in locating the epidural space, it should not be the sole source of information. The most common method is feeling a loss of resistance while injecting fluid with a low-friction glass syringe; however, injecting a small amount of air may be superior because fluid injected in the epidural space may later be aspirated through the needle and give the false impression of being in the subarachnoid space. After multiple passes at one spine level have been performed, the loss-of-resistance method may lose its reliability. Therefore inserting a Seldinger wire through the needle can provide invaluable information about the degree of penetration into the spinal canal. If the needle tip is in the interspinous ligament and has not penetrated the ligamentum flavum, the wire cannot be advanced. The wire can be advanced only if the needle tip is in the paraspinal muscles or within the spinal canal. The pattern of advancement and the location of the wire under fluoroscopic imaging can further clarify its position.

If more than one electrode is inserted, it is wise to insert the other needles before inserting the electrodes since subsequent needle insertion might shear an already implanted electrode. Moreover, it is often possible to insert two electrodes simultaneously and advance them synchronously in the epidural space while maintaining their relative position and spacing.

Once the electrode is in the spinal canal, one has to be certain that it is positioned in the epidural space and not within the subarachnoid space. Although this may seem obvious, it may require multiple attempts at needle placement. This is particularly true if the dura has been previously pierced and CSF has escaped and pooled in the dorsal epidural space. One major difference is that in the subarachnoid space much less resistance is encountered when moving the electrode, particularly for lateral movements. The wire seems to be “floating” and undergoes large shifts of direction in the subarachnoid space; whereas in the epidural space movements are more discrete and obtained only with specific manipulations. Nevertheless, unfortunately the same type of wire/electrode movement can be experienced epidurally if the dural sac has collapsed significantly secondary to CSF loss. In such instances electrical stimulation clarifies the position; a subarachnoid placement elicits responses at much lower thresholds than an epidural placement.

When the epidural space is satisfactorily identified, the electrode is inserted gently under fluoroscopic guidance in the anteroposterior view. Once it has been inserted through the tip of the needle, the electrode must be removed without damage to the insulation surrounding the electrical contacts. If the electrode does not slide without minimal resistance, the needle and the electrode should be removed together. Every time the electrode is withdrawn through the needle, it should be inspected for minute breaks in the insulation, which would necessitate its disposal. Alternatively, a sleeve can be inserted over the guidewire in the epidural space. The guidewire is then removed, and the electrode is inserted through the sleeve. This obviates the risk of shearing the electrode during manipulation.

During placement the electrode may curve around the dural sac to the ventral epidural space. In the anteroposterior projection this might be indistinguishable from a proper midline dorsal location. A gentle lateral curve of the electrode shortly after its entry in the epidural space should raise the suspicion that it is actually lying ventrally. Absolute confirmation of the ventral location arises from stimulation generating violent motor contractions or observation in the lateral plane. An x-ray film should always be obtained to document electrode level and position. Care must be taken to identify either the bottom most visualized rib as the twelfth rib or to use the sacrum to count the number of lumbar vertebra to document the level of implant.

Once in place, the electrode must be secured to the interspinous ligament to minimize dislodgment. An important measure is to secure the wire as loops at the electrode insertion site to relieve the strain and reduce migration during bending. Applying suture directly to the leads has been shown to be safe since it does not cause substantial physical damage or any electrical impairment.⁵⁸ Nevertheless, manufacturers still do not recommend tying directly to the lead and continue to recommend the use of anchors to secure leads. When anchors are used to secure the electrode, one must remember to secure the anchor to both the electrode and the fascia.

Placement of Paddle Electrodes

Paddle electrodes require a surgical procedure for implantation under direct vision.⁵⁹ The patient may be placed in one of two basic positions: prone or semilateral. The prone position allows a more intuitive understanding of the spatial relations, which is more familiar to surgeons. At the same time, it can be difficult to obtain adequate sedation and maintain the airway in this position. In the semilateral position the surgeon has access to the spine and the flank, abdomen, or buttock for the implant of the pulse generator. The patient is asked to place himself or herself in the most comfortable position. If the pain is predominantly on one side, the patient is asked to lie on the less affected side. In this position airway management is safer than in the prone position, and the anesthesiologist is more comfortable in keeping the patient deeply sedated. For cervical placement the neck is slightly flexed but not excessively rotated laterally because extreme rotation substantially increases the difficulty of electrode placement. In general the variable degree of rotation with the patient in the semilateral position makes it more difficult to determine the location of the midline, which may be a significant problem in the cervical area.

A laminotomy is performed approximately one level below the intended level to allow the plate electrode to rest at the planned target for stimulation. Localization is performed with fluoroscopy or with a plain x-ray film, using metallic markers placed on the skin at the level of the planned incision. In a thin patient the incision is about 1 inch in length; and even in large individuals the incision seldom exceeds 2 inches because three to four levels can be accessed by extending the dissection and stretching the skin edges with a Gelpi retractor. Subperiosteal dissection is performed to expose the upper half of the inferior spinous process and the entire superior spinous process. Parts of the superior spinous process are then incrementally removed to expose the interposing ligamentum flavum. In the lower thoracic and upper lumbar area this usually requires the removal of the inferior third of the spinous process. However, in the midthoracic area the acute angle and significant overlapping of the spinous processes necessitates the removal of the whole spinous process. Following removal of the ligamentum flavum, the electrode is inserted into the dorsal epidural space; and its position is confirmed with fluoroscopy and by functional testing.

Such testing previously consisted of waking the patient and determining efficacy of stimulation from patient responses. This has gradually been replaced with intraoperative stimulation with electromyography (EMG) stimulation coupled with somatosensory evoked potential (SSEP) testing, which helps detect proper positioning of the electrode. EMG stimulation is performed with a signal consisting of a greater than 310 microseconds pulse width at a frequency of 5 Hz; the amplitude is gradually increased until EMG signal changes are detected. Alternatively, awake testing is commonly performed with 50-Hz stimulation. As one would expect, bilateral extremity stimulation suggests midline placement, and early root onset implies too lateral of a placement. In addition, stimulation of the dorsal columns results in a greater area of coverage with paresthesia—the percentage of body surface covered is inversely proportional to the distance between two electrodes.⁶⁰ To achieve this, most of the current has to be directed to an area within 2 to 3 mm on each side of the physiological midline. It is important to note that the physiological midline may differ from its anatomic counterpart by up to 2 mm¹; in such instances one should rely on the former since it has a higher clinical correlation. With this feedback, the implant should be placed midline in cases of axial symptomatology or with a bias to one side for patients with unilateral pain. By correlating paddle position to the prior position of trial electrodes and testing motor stimulation responses, one can effectively find the proper location for implantation without waking the patient.

As with percutaneous leads, it is important to secure the leads with a strain relief loop with a diameter no less than 2.5 cm to minimize lead migration with movement. This loop can be sutured directly to the paraspinal muscles before closing the fascial layer.

Implantable Pulse Generator Placement

The IPG should be placed in an area where the unit will have enough overlying soft tissue but also not be pressing against bone, usually the abdomen or buttock. To find the appropriate area of implantation in the latter, three boney prominences are used as landmarks. The posterior superior iliac crest marks the medial border; the greater trochanter of the femur and the apex of the iliac crest define the lateral border. The IPG is implanted approximately 2 to 3 cm deep in the lateral aspect of this triangle (Fig. 5-9).

Fig. 5-9 Landmarks for battery placement.

Postlaminectomy Syndrome/Failed Back Surgery Syndrome

In the treatment of postlaminectomy syndrome, a great challenge has been to obtain stimulation in the low back. Even with direct stimulation to the low back, the pattern of paresthesia is often replaced in time by an unpleasant segmental band of stimulation from the thoracic roots, which negates the benefits of the procedure. Previous pioneering work by Jay Law ^61,⁶² has shown that stimulation in the low back can be obtained only if one uses multiple arrays of closely spaced bipoles at T9-T10. North and associates⁶³ have challenged the concept of the superiority of centered dual electrodes by showing that one single quadripolar electrode in midline has the ability to stimulate the axial low back. These were acute observations, and no data exist as to the long-term behavior of single vs. dual electrodes. The advent of the tripole electrodes and the ability to steer current has made it more plausible to aim for low-back paresthesia. This technology, which consists of a cathode flanked by lateral anodes, also theoretically increases the discomfort threshold.⁴⁶

Angina

The most appropriate electrode location for the treatment of angina pectoris is most likely the lower cervical and upper thoracic region, although some have reported successful higher cervical placements.⁶⁴ In terms of its daily use, most patients use a low-intensity stimulation for several hours per day for prophylactic purposes.⁶⁵

Brachial Plexitis/Neurogenic Thoracic Outlet Syndrome

With the currently available systems, stimulation of all areas is often not feasible with a single electrode. Stimulation of the shoulder and entire upper extremity requires an electrode placement in the upper cervical area; whereas stimulation in the axilla and upper chest wall area is typically achieved only with an electrode in the lower cervical/upper thoracic area. Only on rare occasions does an upper cervical placement reach all of the aforementioned areas. Therefore typically two electrodes are necessary to obtain appropriate coverage. In addition, a two-channel stimulating device might be indicated because the setting for the upper cervical region may be different from that required to optimally stimulate the axilla and anterior chest area.

One specific challenge with these patients is in the extreme hyperesthesia present in the brachial plexus region and the trapezius, which often extends to even larger areas of the posterior thorax. Any surgical manipulations in these areas of hypersensitivity must be avoided at all cost. Tunneling of the implanted wires must be planned carefully and must be at a substantial distance from the hypersensitive area. Lack of adherence to these principles often results in excruciating pain and intolerance to the implanted system.

Patient Management/Evaluation

Careful follow-up of the patients is necessary for successful long-term satisfaction. Equipment-related problems can arise at any time after implantation. Such problems include discomfort at the pulse generator/radio receiver site, electrode breakage or migration, and infection. Titration of the electrical parameters and careful follow-up of the patients is just as important as a technically correct surgical implant procedure. This process can be time consuming. The use of computer-assisted and patient interactive programming is showing great promise;^66,⁶⁷ the ability to download several programs into the pulse generator/receiver, which can later be recalled by the patient at the push of a button, has greatly facilitated the management of the implanted stimulation systems.

There are three very important electrical parameters: the perception threshold, the discomfort threshold, and the usage range. The perception threshold is the voltage where the patient begins to perceive paresthesia;⁶⁸ the discomfort threshold is the voltage where the patient perceives the stimulation as uncomfortable, either because the paresthesias are too strong or because it elicits motor contractions; and the usage range is the difference between these two parameters. As previously indicated, the perception threshold is the lowest in the cervical area and highest in the midthoracic area because of the increasing thickness of CSF dorsal to the spinal cord in the midthoracic spine. For this reason electrode placement in the midthoracic spine favors more selective activation of the segmentary sensory system and seldom results in satisfactory long-term stimulation of the dorsal columns. The distance between electrode and nerve fibers is also affected by patient posture⁶⁹ and thus may require use of different programs throughout the day for such variations. This can be cumbersome, and no technology yet exists that is capable of adapting to these changes.

Once the optimal combination and stimulation voltage have been established, other electrical parameters must also be assessed to optimize the stimulation. The rate of stimulation is very important. Although most patients choose rates between 20 and 100 Hz, there is a high individual variability, and several different rates must be attempted before settling on the optimal one. Some evidence exists that patients with specific neuropathic pain syndromes may respond better to higher stimulation rates (>250 Hz).⁶⁶ The other parameter of pulse width provides a finer adjustment for the current intensity; and other features such as cycling, ramping, and biphasic pulses should be explored to promote the most effective delivery of stimulation. It is not unusual for optimal stimulation parameters not to be fully established for several weeks following implantation.

Outcomes Evidence

In selected patients SCS can produce at least 50% pain relief in 50% to 60% of the implanted patients; and with the proper follow-up care these results can be maintained over several years.

Postlaminectomy Syndrome/Failed Back Surgery Syndrome

Longitudinal studies by North and associates ⁷⁰ evaluated 50 patients with postsurgical lumbar arachnoid or epidural fibrosis without surgically remediable lesions. In this group of patients SCS was superior to repeated surgical interventions on the lumbar spine for back and leg pain and superior to dorsal ganglionectomy for leg pain. Specifically a reduction in pain by at least 50% and patient satisfaction were obtained in 53% of patients at 2.2 years. These results are consistent with a 1995 systematic review of 41 studies,⁷¹ in which approximately 50% to 60% of patients with failed back surgery syndrome (FBSS) had greater than 50% pain relief from SCS. Similarly, Burchiel and associates⁷² conducted a prospective multicenter study in 1996 with 1-year follow-up and also reported 55% successful stimulation; but unfortunately medication usage and work status were not significantly changed. In 2007 a small retrospective series also demonstrated patient satisfaction of 73.5% with reduction in visual analog scale (VAS) scores and amount of additional medication required.⁷³

North, Kidd, and Piantadosi ⁷⁴ also conducted a prospective study randomizing patients with FBSS to repeat back surgery or SCS surgery; but allowing patients to cross over after 6 months. Although 10 of 15 patients crossed over from back surgery to SCS, only two of 12 patients crossed over from SCS to back surgery. North and colleagues⁴⁷ also compared the difference between paddle and percutaneous electrodes. Although laminectomy electrode placement is more invasive than percutaneous placement, it yielded significantly better clinical results in patients with FBSS at up to 3-year follow-up.

The PROCESS trial, published in 2008, involved randomizing 100 patients with FBSS to medical management alone or to SCS in addition to medical management.⁷⁵ In this 2-year study, the authors found that patients randomized to SCS demonstrated >50% leg pain relief, improved quality of life, and greater functional capacity more frequently (37%) than those randomized to medical management alone (2%). However, once again no difference was noted in the amount of opioid use between the two groups.

In a recent prospective cohort study of workers’ compensation recipients who underwent SCS for FBSS, no significant difference was detected with regard to pain, function, medication use, and work time loss after 6 months of follow-up.⁷⁶ This study emphasizes the importance of patient screening to increase chances of a favorable outcome since those who underwent SCS in this study had low average mental health scores, poor baseline functioning, and axial back pain in addition to radicular leg pain.⁷⁷

Complex Regional Pain Syndrome

In 1989 Barolat, Schwartzman, and Woo ⁷⁸ reported reduction of pain in 10 of 13 patients implanted. No patients in that series were made pain free, but all 10 reported a definitive difference when the stimulation was stopped. In 1997 Kumar, Nath, and Toth⁷⁹ presented a median follow-up of 41 months on 12 patients with permanently implanted leads. Eight patients reported near complete resolution of their symptoms, and four also maintained good relief.

Kemler and associates ⁸⁰ reported 23 cases with 78% of the patients reporting improvement. In 2000 this group published another series of 54 patients who underwent randomization to SCS with physical therapy or to physical therapy alone.⁸¹ Using an intent-to-treat analysis, in the SCS-randomized group 67% of patients experienced significant pain relief, which persisted at 6 months. Specifically 2.4 cm decrease in VAS for pain was noted in the SCS group; compared to 0.2 cm increase in the group undergoing physical therapy alone. At the same time, no functional improvement was observed in either group. However, in subsequent publications Kemler and associates⁸² noted that the effects of SCS that were maintained at 2-year follow-up, diminished over time for these patients such that, at 5-year follow-up, results were no longer statistically significant in favor of SCS. Nevertheless an as-treated analysis revealed near statistical significance for pain relief (p = 0.06) and statistically significant improvement in global perceived effect. When interviewed, 95% of patients with an implant reported that they would repeat treatment for the same result.⁸³

Oakley and Weiner ⁸⁴ reported a prospective study of 19 patients with CRPS implanted with SCS systems. Of the 10 patients about whom detailed long-term efficacy data were available, three reported full relief from their pain, and seven reported partial relief.

Three additional prospective studies without matched controls have been reported (total of 50 subjects).^59,^84,⁸⁵ Two of the studies reported success rates with an 84% overall success rate. The third study by Calvillo and colleagues⁸⁵ reported a significant improvement in pain scores (VAS) and a >50% reduction in narcotic use by 44% of subjects. In eight retrospective studies the overall success rate was 84% (192 patients).⁸⁶

Chronic Critical Limb Ischemia (CCLI) and Pain

After the work by Cook and Weinstein,²⁶ in 1981 Meglio and associates⁸⁷ reported pain relief and ulcer healing in a patient with advanced peripheral arterial insufficiency. In 1988 Jacobs and colleagues⁸⁸ published clinical evidence that SCS improved the microcirculation as measured by capillary microscopy.

An early trial was performed by Klomp and associates,⁸⁹ who randomized 120 patients with critical painful limb ischemia to receive medical therapy alone or SCS in conjunction with medical therapy. At a mean follow-up of 19 months there was no significant difference in pain score improvement between the two groups. Yet Jivegard and colleagues⁹⁰ found benefits of SCS when they studied 51 patients who were randomized to receive either oral medication alone or SCS with oral medication. This group reported a significant improvement in pain scores of the SCS-treated group over the non-SCS group (p < 0.01). Similarly, a prospective randomized trial by Guarnera, Furgiuele, and Camilli⁹¹ demonstrated superior pain control and increased ulcer healing of SCS (72%) compared to distal arterial reconstruction (40%).

Four prospective studies without matched controls reveal an overall success rate of 78% (n = 271); and an analysis of seven retrospective studies found an overall success rate of 76% (n = 308).⁸⁶ A review of the European literature demonstrates that 70% to 80% of patients achieved significant (>75%) pain relief. Moreover, pain relief with SCS was found to have more persistent effects than medical treatment alone. Cochrane reviewers have found evidence to favor SCS over conservative treatment to improve limb salvage and achieve pain relief in those unfit for surgery or who have persistent distal ischemia and pain despite revascularization.^92,⁹³

Angina

Vulinkand colleagues ⁹⁴ conducted a prospective study on quality of life changes in patients with refractory angina pectoris implanted with SCS. They found that both the pain and the health aspects of quality of life improved significantly after 3 months of SCS. Further, social, mental, and physical aspects of quality of life were found improved after 1 year of SCS.

Hautvast and associates ⁹⁵ implanted SCS in patients with stable angina pectoris and randomized them. One group remained inactivated, and the other group was instructed to use the stimulator three times per day for 1 hour and with any angina attack. At 6 weeks the treatment group had increased exercise duration and time to angina and decreased angina attacks and sublingual nitrate consumption. Also observed was a decrease in ischemic episodes on electrocardiogram (ECG), with a decrease in observed ST segment depressions on exercise ECG. There was an increase in perceived quality of life and decrease in pain. It was shown that a placebo effect from surgery in the treatment group was unlikely because all patients had implantation surgery at baseline.

In another prospective study of 104 patients who underwent SCS implantation for refractory angina pectoris there was a significant decrease in angina episodes at rest, angina episodes with activity, and total angina episodes.⁹⁶ Five additional studies are reported to be prospective but without matched controls.^94,⁹⁷^–¹⁰⁰ Each of these also revealed significant benefit from SCS. The benefit indices ranged from reduction in angina attacks, decreased nitrate consumption, decrease in New York Heart Association (NYHA) grade, and improvement in Nottingham Health Profile (NHP) grade.

A recent randomized controlled study in patients already implanted with SCS for angina, demonstrated improvement in functional status and symptoms in treatment arms with conventional or subthreshold stimulation in comparison to a low-output placebo treatment arm.¹⁰¹ This is the first blinded study in which stimulation below the sensory threshold for paresthesia has demonstrated therapeutic efficacy while eliminating the possibility of a placebo effect.

To compare SCS to coronary artery bypass graft (CABG), Mannheimer and associates ⁶⁴ randomized 104 patients accepted for CABG to receive either CABG (n = 51) or SCS (n = 53) in the electrical stimulation versus coronary artery bypass surgery in severe angina pectoris (ESBY) study. This study demonstrated that patients randomized to SCS showed a greater than 30% improvement in Nottingham Health Profile (NHP) scores when compared with their baseline. This difference was significant and comparable to that shown by patients randomized to CABG¹⁰² and were consistent on follow-up after 4 years. More important, the 5-year mortality of 27.9% in the ESBY study was similar between those receiving SCS and those who received CABG, with no difference in the percentage of cardiac deaths. In addition, whereas cardiac events were similar across the groups, significantly more cerebrovascular events were observed in the CABG group. Both groups experienced a significant and similar reduction in the number of angina attacks and the consumption of nitrates. Ultimately SCS appears to be equivalent to CABG in terms of symptom relief and may be an alternative to patients with increased surgical risk.

The concern of whether or not stimulation can conceal an acute myocardial infarction was directly addressed by Andersen, Hole, and Oxhoj.¹⁰³ They reported on 3 out of 45 patients treated with SCS for angina pain who survived a myocardial infarction (MI). All three patients noticed the pain to be different and unrelieved with SCS, and all patients correctly guessed that the pain was caused by a myocardial infarction. Therefore it was believed that SCS does not seem to conceal an acute MI.¹⁰⁴ Instead, SCS is believed to reduce the severity of angina attacks but not be able to suppress the conduction and perception of cardiac pain signals that act as indicators of cardiac distress.¹⁰⁵ Similarly, Murray and associates¹⁰⁶ have shown that SCS for refractory angina is effective in preventing hospital admissions without masking the ischemic symptoms or leading to silent infarction.

Abdominal/Visceral Pain Syndromes

Several early studies have since demonstrated the benefit of SCS in abdominal visceral disease. There was an initial hesitancy toward the application of SCS for visceral and somatic pain secondary to the belief that nociceptive pain could not be modulated via stimulation. However, Ceballos and colleagues ¹⁰⁷ demonstrated reduction in pain scores and decrease in narcotic use in a patient treated for mesenteric ischemia with stimulation at T6. Similarly, Krames and Mousad¹⁰⁸ described a patient treated for irritable bowel syndrome who was developing escalating pain and diarrhea. After a stimulator was placed at T8, there was a subjective decrease in pain from 9/10 to 2/10 with only two diarrhea episodes and significant reduction in pain medications in the first 6 months. Even though the patient’s pain recurred after 10 months, the significant reduction in the amount of diarrhea at that time persisted.

Kha, Raza, and Khan ¹⁰⁹ reported the largest series of nine patients with refractory abdominal pain. Five of the nine patients had nonalcoholic pancreatitis, three had a presumed abdominal wall neuroma from frequent abdominal surgery, and another had postsplenectomy pain after trauma. All patients had a significant improvement in VAS scores and decreased narcotic use with placement of the leads at the T5-T7 level at 6- to 8-month follow-up.

Tiede and associates ¹¹⁰ reported on multiple abdominal surgeries and failed conservative measures. Each patient had an element of postprandial abdominal pain associated with nausea and vomiting. In both patients the leads were placed at the T2 level with significant improvement in pain, decreased narcotic use, and increased functioning such as return to work.

More recent studies have looked at the treatment of visceral pelvic pain specifically with SCS applied to the dorsal columns. Kapural and colleagues ¹¹¹ reported on the value of neurostimulation for chronic visceral pelvic pain in six female patients with long-standing pelvic pain secondary to endometriosis, multiple surgical explorations, and dyspareunia. At an average follow-up of 30 months, they found a significant decrease in the VAS score with an average of more than 50% pain relief, accompanied with a decrease in opiate use.

Additional case reports have also demonstrated benefit of SCS for treatment of intractable abdominal or other etiologies, including familial Mediterranean fever and Bannayan-Riley-Ruvalcaba syndrome.^112,¹¹³

Risk and Complication Avoidance

With the proper expertise, permanent complications are rare.⁸⁸ The most serious complication, which is shared with any type of spine surgery, is paralysis or severe neurological deficits, which are more likely during the dissection of epidural adhesions. Fortunately the number of cases of neurological injury involves far less than 1% of implants. An epidural hematoma is another rare but serious complication that can result from venous plexus damage on electrode insertion.

Electrode migration is not an uncommon problem and may require a reoperation to recapture adequate paresthesia. This phenomenon tends to occur most commonly in the first few days since it takes approximately 3 to 4 weeks for scar tissue to form around the plate electrode. In a retrospective review performed in 2006, lead migration was found to be slightly more common with plate electrodes (11.4%) than with percutaneous electrodes (9.8%); and this complication was also found to occur sooner in the former.¹¹⁴ Close attention to programming is essential since a change in voltage requirements or sudden loss in previously captured paresthesia may hint at this problem. The incidence of electrode migration has varied significantly in the literature but is more accurately between 1% and 15%.

As technology continues to improve, hardware-related complications continue to decrease. Overall complications with the equipment include a 1% to 4% incidence of unspecified hardware failure or <2% incidence of battery failure. There is also a 1% to 8% incidence of lead breakage, which is common in the cervical spine and with plate electrodes.¹¹⁴ An x-ray film may not reveal the break—particularly if the break is limited to a small number of filaments—and high impedance is the only indication of this diagnosis. Most hardware-related problems are simply corrected with replacement.

Infection of the implanted hardware has occurred with a 3% to 5% rate. Infection usually affects the IPG or the cabling connecting to the electrode; very seldom does infection spread to the epidural space. Patients typically present with persistent tenderness over the implanted hardware, which can be difficult to differentiate from their baseline pain. The ultimate treatment is complete removal of the hardware followed by a prolonged course of intravenous antibiotics. Even when the infection is limited to the IPG, removal of that component alone leaves the patient at risk for spread of infection to the electrode.

A CSF leak may occur in less than 1% of implants. The clinical presentation consists of spinal headaches and fluid accumulation at the pulse generator site. These patients can be treated initially with bed rest for a few days and a tight abdominal binder over the pulse generator site for 2 to 3 weeks. If symptoms persist, a blood patch or surgical exploration may be required. When the CSF leak threatens the incision, surgical intervention is imperative.

Persistent pain at the implant site has been seen in about 5% of patients with CRPS. Recalcitrant CSF leakage has been encountered in a few patients, requiring multiple surgical revisions. Breakage or malfunction of the implanted hardware, particularly the electrodes and the subcutaneous extension cables, has been encountered in about 10% of the implanted systems. Disruption of the programmed settings can be minimized with appropriate patient education. For example, patients should be informed to turn off the device before going through an antitheft device in a retail store and to avoid strong magnetic fields, including magnetic resonance imaging (MRI) scanners. Painful stimulation, necessitating either repositioning or removal of the electrode, has also been reported in a number of cases.