Establishing Diagnosis

Spinal cord stimulation (SCS) is an adjustable, nondestructive, mode of neuromodulation that delivers therapeutic doses of electrical current to the spinal cord for the management of neuropathic pain. The most common indications include postlaminectomy syndrome, complex regional pain syndrome (CRPS), ischemic limb pain, and angina (Table 5-1); but it has been applied to other causes of intractable pain.

Table 5-1 Spinal Cord Stimulation Indications and Probability of Success

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 pattern¹ 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 Wall⁶ 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.

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 associates⁴⁷ 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.