Chapter 2 Mechanisms of Spinal Neuromodulation
Chapter Synopsis: Electrical stimulation of the spinal cord (SCS) improves many forms of neuropathic pain; but, contrary to our early understanding, it can also affect some forms of nonneuropathic nociception. Chapter 2 examines the physiology of these indications. The understanding of SCS is rooted in Melzack and Wall’s5 gate control theory of pain transmission. By spinal stimulation of large-fiber neurons, the gate is activated to reduce transmission of neuropathic pain signals from primary small-fiber afferents. The technique generally does not alleviate acute nociception, but it can reduce certain types of peripheral nociception and can even alleviate underlying conditions. SCS has been shown to affect ischemic limb pain caused by peripheral arterial occlusive disorder (PAOD), angina, and gastrointestinal disorders such as irritable bowel syndrome (IBS).
Transmission of nociceptive information from the site of an injury may generate a perception of pain because an imbalance exists between large and small fiber systems (cf. Head and Thompson1). This concept eventually evolved into Melzack and Wall’s5 gate control theory.
Background
Therapeutic effects of neuromodulation are based on the concept that selective excitation of large afferent fibers activates mechanisms that control pain. This fits well with the idea that pain may occur as a result of an imbalance between large and small fiber systems that transmit nociceptive information from the site of injury. Previous investigators have provided a long history of support for this concept. As early as 1906 Head and Thompson1 argued that fine discrimination such as touch normally exerts an inhibitory influence on impulses transmitted in fibers mediating nociception, which results in pain. This inhibition or facilitation of sensory impulses has been proposed to occur in the dorsal horn before nociceptive information is relayed onto secondary neurons. Furthermore, clinical trials performed in the early sixties using sensory thalamic stimulation2,3 were based on the notion that activation of fine discrimination receptors (touch) exerted an inhibitory influence over sensations such as pain, pressure, heat, or cold. It should also be noted that Noordenbos4 used the descriptive phrase “fast blocks slow” to stress the inhibitory influence of fast on slow fibers.
The concept of excitation of large afferent fibers activating pain control mechanisms advanced very rapidly with the publication of the article proposing the gate control theory; it is one of the most cited papers in modern pain literature.5 In this article the authors suggested that the therapeutic implication of their model would be to selectively activate large fibers to control pain. Thus even though the basic idea underlying the gate control theory was not completely unknown, it was built on a foundation of creative experiments using modern electrophysiological techniques. The results of these experiments were clearly synthesized and discussed in a form that postulated a new conceptualization of pain and pain control. Subsequently, numerous studies were conducted to criticize the theory, but nevertheless its simplicity has provided a useful frame of reference to explain mechanisms of pain generation and pain control. As Dickenson6 pointed out in his editorial about the ability of the gate control theory of pain to stand the test of time, the concepts of convergence and modulation changed the focus from destructing pathways for relief of pain to controlling pain by modulation in which excitation is reduced and inhibition is increased. The gate control theory accelerated the pursuit of modern pain research to explore how the pervasive plasticity of the nervous system plays a critical role in the generation, maintenance, and modulation of pain.
The gate control theory served as a critical catalyst in the clinical arena to spawn the development of various forms of neuromodulation that led to new therapies. The insights gained by Shealy and his colleagues7 and Shealy, Mortimer, and Reswick8 in animal experiments led them to conduct the first human trials with electrical spinal cord stimulation (SCS) as one form of neuromodulation.8 Their experimental studies in conscious cats revealed that stimulating the dorsal aspect of the spinal cord blocked responses to nociceptive peripheral stimuli. On the basis of this study and support of the gate control theory, it was assumed that neuromodulation could be used to treat all forms of nociceptive pain. However, several reports pointed out that SCS is ineffective for treating acute nociceptive conditions in contrast to what was predicted from the gate control theory; but eventually it has become the foremost treatment for neuropathic pain originating from the periphery.9–13 Nevertheless, numerous reports appeared during the eighties to convince clinicians that SCS could also be used to alleviate certain types of nociceptive pain, including selected ischemic pain states such as peripheral arterial occlusive disease (PAOD), vasospastic conditions, and therapy-resistant angina pectoris. The mechanisms of action for SCS are slowly emerging as more solid evidence has revealed some of the underlying physiological mechanisms. Clinical observations coupled with important experimental data clearly demonstrate that SCS applied to different segments of the spinal cord elicits fundamentally different results on various target organs or parts of the body (Fig. 2-1).
Organization and Electrical Properties of the Spinal Cord
The spinal cord is encased within the vertebral canal, which is made up of vertebrae that encircle the spinal cord but limit space for insertion of stimulating electrodes. The spinal cord in an adult human extends from the foramen magnum to the first or second lumbar vertebra and is divided into cervical, thoracic, lumbar, and sacral segments. The naming of the segments is based on the regions of the body innervated by the spinal cord. Examination of a cross section of the spinal cord shows that it is composed of gray matter and surrounded by white matter (Fig. 2-2).
The gray matter is comprised of cell bodies with their dendrites and initial segment of the axon, microglia, and astrocytes. It is divided into a posterior horn, intermediate zone, and the ventral horn. The gray matter is further divided into laminae I to X; these divisions are based on the size, shape, and distribution of neurons located in these laminae.14 The input received by these neurons and the trajectory of the axons from them also help to characterize laminae. Neurons of dorsal and intermediate laminae (I to VII, X) generally receive sensory information originating from peripheral sensory receptors. These neurons integrate this information with input arriving from descending pathways. Some of the cell bodies have short axons and serve as interneurons, whereas others are the cells of origin of ascending sensory pathways. The interneurons may also participate in local reflexes. The ventral laminae (VIII, IX) are generally composed of motoneurons that form the motor nuclei.
The electrical properties, more specifically the electrical conductivity, of white and gray matter of the spinal cord are not homogeneous. For SCS it is important to know that the electrical conductivity of the dorsal column is anisotropic; that is, current can travel in the direction parallel to the axons more easily than in the direction perpendicular to axons.15 The electrical properties within the gray matter also vary because neurons and glia have diverse orientations, ubiquitous dimensions, and different dendritic characteristics.
Neuromodulation using electrical stimulation of the spinal cord depends on the conductivity of the intraspinal elements relative to the position of the electrode.16 If an axon is depolarized or made more electrically positive, it produces an action potential that is transmitted orthodromically and antidromically within the axon. The cathode of an external electrode must be negatively charged to generate the action potential in the axon. In contrast, if an axon is hyperpolarized or made more negatively charged, its ability to generate an action potential is reduced because the threshold for depolarization is increased. A positively charged external electrode or anode produces this effect. Thus the active electrode for electrical stimulation serves as the cathode, whereas the anode or positive electrode may serve as a shield to prevent stimulation of neuronal structures such as dorsal roots that might interfere with effective neuromodulation. For SCS the electrode most commonly is placed on the surface of the dura mater. Activation of the electrode releases electric current that is transmitted through the dura mater and the highly conductive cerebrospinal fluid (CSF) before it reaches the dorsal part of the spinal cord. The dura mater has low conductivity, but it is so thin that the current generally is not impeded significantly as it passes through the dura to the CSF. Furthermore, the vertebral bone has the lowest conductivity so it insulates pelvic structures and visceral organs from the electric field generated by SCS. Once the electric current reaches the spinal cord, several factors may determine the neural structure being stimulated. Jan Holsheimer16 has used computerized models of the spinal cord to study the activation of axons by electrical current. In addition to the fiber diameter, the presence of myelination, and the depth of CSF layer surrounding the cord at the level of an electrode, the axon orientation has important implications for activation thresholds. In general, axons of the dorsal columns have higher activation thresholds than fibers such as the dorsal roots that are oriented laterally or angle as they enter the spinal cord.16
The dorsal column is composed primarily of large-diameter afferent nerve fibers with relatively low thresholds for recruitment when cathodal electrical pulses are generated through the epidural electrode that is attached to a spinal cord stimulator. It is important to note that the electrode for SCS needs to be placed near midline to prevent the activation of dorsal root fibers.17 Stimulation amplitudes are then increased to intensities that recruit large fibers to produce action potentials and produce paresthesias. These action potentials are transmitted orthodromically and antidromically in these axons. The action potentials transmitted antidromically reach the collateral processes that penetrate the gray matter of the spinal cord. Their activation causes the release of transmitters, which activates the “gate.” Activation of the gate sets in motion neural mechanisms that reduce pain and improve organ function. The details of these mechanisms are discussed in subsequent paragraphs.
Neuromodulation Mechanisms in Ischemic Pain
Ischemic painful conditions of the limbs commonly results from PAOD, which is caused by obstruction of blood flow into an arterial tree.18 PAOD is a major cause of disability and loss of work and affects the quality of life.19,20 Morbidity and mortality are relatively high because effective treatments are very limited. Presently SCS is usually implemented only after vascular surgery and medications fail to slow or prevent the progression of PAOD. Surprisingly the success rate of SCS-treated PAOD is greater than 70%.21 Since ischemic pain is characterized generally as essentially nociceptive and several studies have indicated that SCS does not alleviate acute nociceptive pain,9,22,23 SCS-induced pain relief is most likely secondary to attenuation of tissue ischemia that occurs as a result of either increasing/redistributing blood flow to the ischemic area or decreasing tissue oxygen demand.24,25 Cook and associates26 were the first to report that SCS increased peripheral circulation of patients suffering from PAOD. Usually SCS is applied to the dorsal columns of lower thoracic (T10-T12) and higher lumbar spinal segments (L1-L2) to increase peripheral circulation in the legs of PAOD patients.
The mechanisms of SCS-induced vasodilation in the lower limbs and feet are not yet completely understood. Since no animal models of PAOD that generate ischemic pain have emerged, normal anesthetized animal models have been used to investigate the physiologic mechanisms of SCS-induced changes in peripheral blood flow (see reference 23 for review). Cutaneous blood flow and calculated vascular resistance in the glabrous skin of ipsilateral and contralateral hindpaws have been determined most commonly by using laser Doppler flowmetry. A thermistor probe placed next to the laser Doppler probe on the plantar aspect of the foot has been used to measure skin temperature. Various interventions such as injections of hexamethonium, administration of adrenergic agonists and antagonists, sympathetic denervation, dorsal rhizotomies, calcitonin gene-related peptide (CGRP) antagonists, nitric oxide synthetase inhibitors, and local paw cooling have been used to explore the underlying mechanisms of peripheral microcirculation. Studies using Doppler flowmetry and interventions for more than 30 years of clinical and basic science studies have resulted in the evolution of two theories to explain the mechanisms of SCS-induced vasodilation. One theory is that SCS decreases sympathetic outflow and reduces the constriction of arterial vessels27,28; the alternative theory is that SCS antidromically activates sensory fibers, which causes the release of vasodilators29, 30 (Fig. 2-3). The theory for SCS-induced suppression of sympathetic activity was based on results from clinical observations showing that a sympathetic block or sympathectomy produced pain relief and vasodilation imitated effects of treatment with SCS.31,32 This theory was tested in animal models in which SCS-induced cutaneous vasodilation in the rat hindpaw at 66% of motor threshold was abolished by complete surgical sympathectomy.33 SCS-induced vasodilation was markedly attenuated after administrating the ganglionic blocker, hexamethonium, or the neuronal nicotinic ganglionic blocker, chlorisondamine. These results led to the suggestion that efferent sympathetic activity, including nicotinic transmission in the ganglia and the postganglionic α1-adrenergic receptors are suppressed by SCS (see Fig. 2-3). The alternative theory of SCS-induced antidromic activation of sensory fibers was confirmed in studies showing that sensory afferent fibers are important for SCS-induced vasodilation and that at higher, but not painful, SCS intensities C-fibers may also contribute to the response30,34,35 (see Fig. 2-3). Thus SCS applied at the spinal L2-L5 segments excites dorsal column fibers that antidromically activate interneurons, which subsequently stimulate spinal terminals of transient receptor potential V1 (TRPV1) containing sensory fibers, which are primarily made up of C-fiber axons.36,37 These fibers transmit action potentials antidromically to nerve endings in the hindlimb. The action potentials evoke mechanisms that release vasodilators, including the most powerful vasodilator, CGRP, which binds to receptors on endothelial cells. The activation of these receptors leads to production and subsequent release of nitric oxide (NO), which results in relaxation of vascular smooth muscle cells (see reference 30 for review). The overall result is that relaxation of vascular smooth muscle cells decreases vascular resistance and increases peripheral blood flow. It should be noted that SCS applied at 500 Hz significantly increased cutaneous blood flow and decreased vascular resistance when compared to the responses induced at 50 Hz and 200 Hz; the effects at all of these frequencies depend on the activation of TRPV1-containing fibers and release of CGRP.38 The clinical use of such findings remains to be determined.
The level of sympathetic nervous system activity may shift the balance between the effects of sympathetic efferent suppression and antidromic activation of sensory afferent fibers. Cooler skin temperatures increase sympathetic activity. A notable observation is that SCS-induced vasodilation of a cooled hindpaw (<25° C) generated an early phase of vasodilation via sensory afferent fibers and a late phase via suppression of the sympathetic efferent activity.39 However, only sensory afferent activation occurred if SCS-induced vasodilation was performed in a warm paw (>28° C). Thus the balance of these two mechanisms most likely depends on the activity level of the sympathetic nervous system. Furthermore, another study showed that preemptive SCS increased the survival rate of skin flaps that were made ischemic by occluding the blood supply to the tissue for as long as 12 hours.40 Concomitant administration of the CGRP-1 receptor antagonist CGRP 8-37 markedly attenuated the cytoprotective effect40
