Mode of Action of Spinal Cord Stimulation in Neuropathic Pain

Transcription

1 S6 Journal of Pain and Symptom Management Vol. 31 No. 4S April 2006 Special Article Mode of Action of Spinal Cord Stimulation in Neuropathic Pain Björn A. Meyerson, MD, PhD and Bengt Linderoth, MD, PhD Section of Stereotactic and Functional Neurosurgery, Department of Neurosurgery, Karolinska University Hospital, Stockholm, Sweden Abstract Spinal cord stimulation emerged as a spin-off from the classical gate-control theory, which, however, does not suffice to explain its clinical effects. Whether or not nociceptive forms of pain may be attenuated remains a controversial issue. Previous experimental studies aiming at elucidating the underlying mechanisms were performed on intact, anesthetized animals and were therefore of limited clinical relevance. Not until recent years have some data on the mode of action accumulated providing evidence that gamma-aminobutyric acid (GABA) ergic as well as adenosine-related mechanisms are involved in the pain amelioration. It appears that the effects are exerted mostly via segmental spinal levels, but recent evidence suggests that a supraspinal loop may also be of importance; this issue remains to be resolved. It should be emphasized that most experimental data pertaining to the mode of action are derived from so-called animal models of neuropathic pain. However, caution must be exercised in the translation of such data from bench to bedside, because some behavioral signs interpreted as pain in such models may be misleading. We still need animal studies to generate basic data but these findings should also be confirmed in humans. J Pain Symptom Manage 2006;31:S6--S12. Ó 2006 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Neuropathic pain, electrical stimulation, rat pain models, GABA, CRPS, neuromodulation The issue of why and how spinal cord stimulation (SCS) may effectively alleviate chronic neuropathic pain has attracted relatively little interest, in spite of the fact that this treatment modality has been in use for more than three Studies related to the topic of this article have been partially supported by grants from the Karolinska Institutet and from Medtronic Europe SA. Address reprint requests to: Björn A. Meyerson, MD, PhD, Department of Neurosurgery, Karolinska University Hospital, SE Stockholm, Sweden. bjorn.meyerson@karolinska.se. Accepted for publication: November 15, Ó 2006 U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved. decades and no less than about 14,000 new cases annually are estimated to be subjected to spinal electrode implantation worldwide, the great majority for neuropathic pain. It is not until recent years that experimental and clinical data have accumulated supplying a base for some understanding of the mechanisms that may be involved in the production of pain relief However, only a few experimental studies have been performed on humans with the aim of exploring the possible modulatory effect of SCS on somatosensory function, in particular pain, and no such investigations have appeared in the last decade. 4 Most of our knowledge therefore derives /06/$--see front matter doi: /j.jpainsymman

2 Vol. 31 No. 4S April 2006 Mechanisms of Spinal Cord Stimulation S7 from animal experiments. In the 1970s and 1980s, many such studies were reported but all of them had been performed on normal, anesthetized animals and, at variance to the clinical characteristics, a possible effect on responses to acute, nociceptive peripheral stimuli of SCS applied with high amplitude for short periods of time was monitored. With such experimental designs, the clinical relevance of animal studies may be questioned though they have to some extent provided background data, which are of value for the interpretation of results from subsequent studies on animal models of chronic neuropathic pain. It is well known that SCS was a direct spin-off from the gate-control theory 5 and this is in fact a very illustrative example of what is presently referred to as translation from bench to bedside. However, if the effects of SCS in patients were produced via the mechanisms implied in the theory it would be expected that SCS should be particularly effective in suppressing acute, nociceptive pain (e.g., ulcer pain, postoperative pain, etc.). One may therefore allege that it is a paradox that the main concept of the gate-control, i.e., the inhibition of transmission of nociceptive signals, which was the base for the development of SCS, cannot be reproduced in its clinical applications. Anecdotal observations already reported in the first clinical studies of SCS indicated that the treatment was inefficacious for chronic and acute as well as experimentally induced nociceptive pain. Thus, in a patient who was effectively treated for chronic sciatic pain, a new pain resulting from a fracture of the same leg was not alleviated by the stimulation. 6 However, in clinical practice the question of whether SCS may, at least to some extent, alleviate nociceptive pain still appears to be a controversial issue. In many published series of patients subjected to SCS, low back pain is a common indication and by many also regarded to be the best. 7 Most authors, however, contend that SCS predominantly influences the radiating pain component or pain in the leg, which represents lumbosacral rhizopathy, 8 but there are also those who claim that the pain component confined to the axial, lumbar portion of the back may be alleviated It is most probable that this axial pain is essentially nociceptive but relevant studies fail to supply detailed mappings of the stimulation-induced paresthesiae, confirming that it is possible with this form of stimulation to provoke paresthesiae also in midline structures. The possible involvement of facet nerve branches as a cause of a neuropathic pain component in these conditions remains to be elucidated. In the first clinical reports on SCS, it was stated that the threshold of induced cutaneous pain was not elevated, 6 and in 1975 Lindblom and Meyerson 12 studied the perception of cutaneous mechanical pain produced by the use of a calibrated flat forceps in patients undergoing SCS treatment. The stimulus was applied in regions both inside and outside the field of SCS-produced paresthesiae. The thresholds were significantly increased only in sites displaying hyperalgesia and allodynia. In normal skin there was no effect on the thresholds, though the painful stimulus was applied in an area covered by the paresthesiae. Identical differential effects were recorded on thermal sensibility assessed by Quantitative Sensory Testing. Although it is universally accepted that the presence of paresthesiae, indicating the activation of the dorsal columns (DC), is a prerequisite for pain relief, it has also been argued that the tingling and vibratory sensations are merely epiphenomena. The therapeutic effects should instead be exerted via the activation of other pathways than the DC. A major candidate as stimulation target has been the dorsolateral funiculus (DLF), but this is unlikely considering that pain associated with extensive deafferentation or direct injury of the DC fibers fails to respond to SCS. Moreover, it is known that preservation of somatosensory responses evoked from the painful region is, as a rule, a prerequisite for a positive SCS effect. Moreover, the DLF fibers running rostro-caudally at a distance from the SCS lead overlying the DCs conceivably have higher activation thresholds than the root fibers entering horizontally. Thus, these latter would primarily be activated generating segmental paresthesiae at the level of the active electrodesdand this only rarely occurs. 13 The possibility that SCS may inhibit also nociceptive input at segmental spinal level has gained some experimental support by the finding that the stimulation may depress a nociceptive flexor reflex in patients. 14 Electrical stimuli applied to the sural nerve territory

3 S8 Meyerson and Linderoth Vol. 31 No. 4S April 2006 induce a contraction of the biceps femoris when the intensity of the stimulation is perceived as a pricking pain sensation. This flexor response conceivably represents the activation of Ad afferent fibers. It has been demonstrated that this reflex may be attenuated by SCS. The effect seems to relate to the clinical pain relieving effect and has been used as an objective correlate to the pain relieving effect of SCS. This is, however, difficult to explain in view of the fact that SCS does not otherwise influence novel acute pain nor evoked, experimental pain resulting from Ad-fiber activation. Neurophysiological Mechanisms Studied in Animal Models of Neuropathic Pain Modern experimental pain research has made extensive use of animal models. The question of the clinical relevance of data obtained from animal models is crucial and when it relates to pain, in particular chronic pain, it is problematic and an issue of debate between clinicians and basic scientists. For obvious reasons, the presence of pain in an awake animal can only be inferred by interpretation of behavioral signs, spontaneous or evoked. This is further complicated by the fact that alteration of behavior as a result of a lesion produced to mimic a certain pain condition may also be associated with other forms of dysfunction than chronic pain. 15 Rats subjected to injury of the sciatic nerve, its branches or the corresponding nerve roots are often referred to as models of neuropathic pain. However, these animals as a rule behave normally, they gain weight, they groom, and they exhibit normal sociability. Their locomotion is normal apart from some motor deficits in the form of characteristic changes of the hind paw posture. These are present regardless of whether or not the plantar surface of the paw is hypersensitive (allodynic) to mechanical and thermal stimuli. Therefore, the posture changes are more probably due to motor axon injury and not primarily and necessarily related to the presence of pain as has previously been proposed. 16 Thus, one cannot infer that these animals suffer from spontaneous, ongoing pain but that their exaggerated responses to peripheral stimuli conceivably mimic evoked pain. Only in rare cases of very pronounced hypersensitivity can shaking and lifting of the paw, without being peripherally stimulated, occasionally be observed, and this could be a sign of spontaneous pain. 17 Considering that in humans, injury to peripheral nerves only occasionally gives rise to evoked and/or spontaneous pain, it is not unexpected that experimental nerve injury in rats does not always result in painful hypersensitivity. Furthermore, chronic neuropathic pain due to nerve injury in patients is associated with cutaneous, mechanical hypersensitivity in only 20%--40% of cases. 15,18,19 A number of studies have been performed in our laboratory using different varieties of models of mononeuropathy, i.e., rats with injury of the sciatic nerve or its branches resulting in hind paw hypersensitivity. 3 In these animals, a miniaturized SCS system has been implanted and the effect of stimulation on evoked pain can be monitored in the awake, freely moving animal. It has been demonstrated that in some of the rats SCS may effectively suppress the allodynia in a way that mimics what can be observed in patients. Thus, SCS applied for 20 minutes with stimulus parameters similar to those used clinically may lead to a significant elevation of the abnormally low withdrawal threshold to innocuous mechanical stimuli (von Frey filaments) and this effect outlasts the SCS by up to 1 hour. There is much evidence that the phenomenon of tactile allodynia is mediated mainly via low-threshold Ab-fibers and that it represents a central state of hyperexcitability. 20 The plasticity changes in the spinal cord following peripheral nerve injury are manifested by persistent augmented responsiveness and a high degree of spontaneous discharge of primarily wide-dynamic range dorsal horn neurons. In acute experiments we have demonstrated that SCS may induce a significant and long-lasting inhibition of both the afterdischarges and the exaggerated principal response in such neurons in nerve lesioned rats. 21 In the clinical setting, this suppression of dorsal horn neuronal activity could correspond to the beneficial effect of SCS not only on the allodynia but also on the spontaneous neuropathic pain. These observations suggest that SCS may selectively influence Abfiber-related functions. This notion is further

4 Vol. 31 No. 4S April 2006 Mechanisms of Spinal Cord Stimulation S9 supported by the finding that the threshold of the early component of the flexor reflex, which is Ab-fiber mediated, is elevated whereas the late C-fiber-dependent phase was found to be unaffected. 22 However, it has also been reported that the C-fiber flexor reflex can be significantly attenuated, but this observation was made in normal, intact animals. 23 Recently, the effect of SCS on long-term potentiation (LTP) has been investigated. 24 It was found that the duration of the LTP response to C-fiber activation was significantly decreased, from about 6 hours to about 30 minutes. It should be noted that in these experiments only the sensitized C-fiber response was influenced while the normal C- and Ab-functions were unaffected. The mechanisms involved in the phenomenon of cutaneous hypersensitivity and ongoing pain as a result of nerve injury are incompletely understood, and the emphasis on large, low-threshold fiber-related functions as pivotal for explaining the effect of SCS is necessarily an oversimplification. It might well be that the mode of action of SCS relates more to, for example, sensitized or awakened nociceptors, a generalized state of central sensitization, descending spinal facilitation, etc. The conceptual basis for SCS presupposes antidromic activation of ascending dorsal column fibers and this implies that the region of action is segmental. Our experimental data support this interpretation but a research group in Beiruth has provided some evidence that instead the major effect might be exerted via a supraspinal loop. 25 What is Known about Transmitter Mechanisms Involved in SCS? For obvious reasons, the application of electric current onto the dorsal aspect of the spinal cord activates a host of transmitter/receptor systems and little is known about which ones are critically involved in the attenuation of chronic, neuropathic pain. Human data from analyses of lumbar cerebrospinal fluid (CSF) in conjunction with SCS are sparse and inconclusive. However, it appears that opioid mechanisms conceivably are not involved. There is some evidence that the substance P (SP) content in human CSF and the spinal release of SP and serotonin in cats tend to increase as a result of SCS. 26,27 However, it might well be that the SCS-induced changes are not necessarily related to its pain relieving effect. In a series of acute experiments using microdialysis in the dorsal horn of nerve lesioned rats, we have demonstrated that SCS reduces the release of excitatory amino acids (glutamate, aspartate) and at the same time the gamma-aminobutyric acid (GABA) release is augmented. 28 It is of special interest that this effect on the GABA system occurred only in rats that in preceding experiments had been found to respond to SCS with significant suppression of hind paw hypersensitivity. 29 These results confirm earlier observations that the state of central hyperexcitability manifested in the development of allodynia after peripheral nerve injury relates to dysfunction of the spinal GABA systems, and it appears that SCS may act by restoring normal GABA levels in the dorsal horn. These findings were supplemented by behavioral experiments where we showed that that the allodynia-suppressive effect of SCS could be counteracted by intrathecal injection of a GABA B antagonist whereas the GABA A antagonist bicuculline was less effective. Conversely, intrathecal administration of GABA or a GABA B agonist, baclofen, markedly enhanced the effect of SCS. 30 This combined effect of the drug and SCS could also be counteracted by a GABA B receptor antagonist. In subsequent studies it was found that rats that were nonresponders to SCS, i.e., their hind paw mechanical allodynia was not attenuated, could be converted to responders with intrathecal administration of low, by themselves ineffective, doses of baclofen. The same potentiating effect was found with adenosine and it can thus be concluded that both the GABAand the adenosine-related systems are directly involved in the pain relieving effect of SCS. 28,31 These results initiated a clinical study where it was demonstrated that it was possible to enhance the SCS effect by simultaneous intrathecal administration of baclofen in low doses. 32 Again, this is a good example of direct transfer of results from the bench to bedside application. Later it was demonstrated that gabapentin, pregabalin, and clonidine may also have similar potentiating effects in nonresponding rats. 33,34 In particular, the results obtained

5 S10 Meyerson and Linderoth Vol. 31 No. 4S April 2006 with clonidine are of interest since it is known that the antinociceptive effect of this substance may relate to an interference with the spinal cholinergic system 35 and if so, the effect of SCS might act also via involvement of these mechanisms. It is a general experience that 30%--40% of well-selected neuropathic pain patients do not benefit adequately from SCS, in spite of the fact that stimulation-induced paresthesiae cover the painful area. The further development of adjunctive pharmacotherapy may be a possible way of enhancing the SCS efficacy in these patients. Fig. 1 depicts a tentative scheme of some essential features of the mode of action of SCS when applied for neuropathic pain. This conceptualization of SCS is necessarily incomplete, in particular with regard to the possible involvement of transmitter/receptor mechanisms. Further, it is primarily based on experiments performed on animal models of mononeuropathy, and as commented on above, such data should be interpreted with great caution. One of the major problems with animal models of chronic pain is that the repertoire of behavioral signs of or responses to nociception are limited and biochemical alterations may be difficult to relate directly to ongoing or evoked pain. On the other hand, translational pain research, which implies a reciprocal approach between bench and bedside, presents as the most effective way of promoting the further advancement and refinement of treatments. 36 This, however, calls for the development of better animal models that more adequately mimic specific pain conditions as well as investigations to confirm animal findings in human experimental and clinical studies. Pain with Dysautonomia Some pain conditions are associated with signs of dysautonomia and they may be either sympathetically maintained (SMP) or sympathetically independent (SIP) (discussion, see ref. 37 ). The most typical example of such pain is Complex Regional Pain Syndrome (CRPS) of both types, which is known to Fig. 1. Schematic representation of the possible mode of action of SCS in neuropathic pain, based on present knowledge derived predominantly from experiments performed on animal (rat) models of mononeuropathy (ref., 3 with permission). Both segmental and supraspinal mechanisms are represented. Possible supraspinal relays are not included because of insufficient knowledge about the organization of a proposed supraspinal loop. The broken arrow lines represent antidromic and full line arrows ortodromic activation in the DC, their collaterals and in primary A-afferents. The diagram does not depict a possible SCS activation of the DLF. It is conceivable that numerous transmitters and modulators are involved in the modulation exerted by interneurons (represented by X ). Descending control of second order neurons is here represented as inhibitory only although there is evidence suggesting also a facilitatory supraspinal input (SPdsubstance P; EAAdexcitatory amino acids [glutamate, aspartate]).

6 Vol. 31 No. 4S April 2006 Mechanisms of Spinal Cord Stimulation S11 have a relatively high likelihood of responding to SCS. 38 There are several studies demonstrating that the efficacy of SCS is related to the temporary outcome of sympathetic blockades, i.e., stellate, lumbar sympathetic blocks, or regional guanethidine blocks. 39 However, there are also studies reporting that also patients who have failed to respond to sympathetic block may benefit from SCS. Thus, in a small series of children, girls years of age suffering from typical CRPS type I, sympathetic blocks had been of no avail but SCS provided effective pain amelioration eventually resulting in complete recovery in most of the patients. 40 For these reasons, the proposed model of SCS mode of action, predominantly based on experimental animal data and illustrated in Fig. 1, does presumably only partially apply to pain condition like CRPS; it should be noted that there are no reports of SCS applied to animal models of such pain. It is known that SCS may produce peripheral vasodilatation and there is experimental evidence that this may be due to an inhibitory effect on SMP vascular tone. 41 However, there might also be that the effect on peripheral vasculature is mediated via antidromic activation of primary afferents and subsequent release of calcitonin generelated peptide (CGRP). 42,43 On the other hand, the presence of a vasodilatory SCS effect is not a prerequisite for pain relief in CRPS 44 but it is not known how this relates to the SIP and SMP variants of CRPS, respectively. Thus, it can be concluded that the mode of action in pain conditions with dysautonomia is very complex and poorly understood. This again illustrates the urgent need of further research on pain mechanisms and how they can be modulated. References 1. Linderoth B, Foreman RD. Physiology of spinal cord stimulation: review and update. Neuromodulation 1999;2: Linderoth B, Meyerson BA. Spinal cord stimulation. Mechanisms of action. In: Burchiel K, ed. Pain surgery. New York: Thieme, 2000: Meyerson BA, Linderoth B. Spinal cord stimulationdmechanisms of action in neuropathic and ischaemic pain. In: Simpson BA, ed. Electrical stimulation and relief of pain, Pain research and clinical management, Vol. 15. Amsterdam: Elsevier B.V., 2003: Linderoth B, Meyerson BA. Dorsal column stimulation: modulation of somatosensory and autonomic function. Semin Neurosci 1995;7: Melzack R, Wall PD. Pain mechanisms: a new theory. Science 1965;150: Nashold BS, Somjen G, Friedman H. Paresthesias and EEG potentials evoked by stimulation of the dorsal funiculi in man. Exp Neurol 1972;36: Barolat G. A prospective multicenter study to assess the efficacy of spinal cord stimulation utilizing a multi-channel radio-frequency system for the treatment of intractable low back and lower extremity pain. Neuromodulation 1999;2: North RB, Kidd DH, Zahurak M, James CS, Long DM. Spinal cord stimulation for chronic, intractable pain: experience over two decades. Neurosurgery 1993;32: Barolat G, Oakley JC, Law JD, et al. Epidural spinal cord stimulation with a multiple electrode paddle lead is effective in treating intractable low back pain. Neuromodulation 2001;4: Ohnmeiss DD, Rashbaum RF. Patient satisfaction with spinal cord stimulation for predominant complaints of chronic, intractable low back pain. The Spine J 2001;1: North RB, Kidd DH, Olin J, et al. Spinal cord stimulation for axial low back pain: a prospective, controlled trial comparing dual with single percutaneous electrodes. Spine 2005;30: Lindblom U, Meyerson BA. Influence on touch, vibration and cutaneous pain of dorsal column stimulation in man. Pain 1975;1: Feirabend HK, Choufoer H, Ploeger S, Holsheimer J, van Gool JD. Morphometry of human superficial dorsal and dorsolateral column fibres: significance to spinal cord stimulation. Brain 2002; 125: Garcia-Larrea L, Sindou M, Mauguière F. Nociceptive flexion reflexes during analgesic neurostimulation in man. Pain 1989;39: Hansson P. Difficulties in stratifying neuropathic pain by mechanisms. Eur J Pain 2003;7: Bennett GJ, Xie YK. A peripheral mononeuropathy in rat produces disorders of pain sensation like those seen in man. Pain 1988;33: Hogan Q, Sapunar D, Modric-Jednacak K, McCallum JB. Detection of neuropathic pain in a rat model of peripheral nerve injury. Anesthesiology 2004;101: Martin C, Solders G, Sonnerborg A, Hansson P. Painful and non-painful neuropathy in HIV-infected patients: an analysis of somatosensory nerve function. Eur J Pain 2003;7:

7 S12 Meyerson and Linderoth Vol. 31 No. 4S April Otto M, Bak S, Bach FW, Jensen TS, Sindrup SH. Pain phenomena and possible mechanisms in patients with painful polyneuropathy. Pain 2003;10: Woolf C, Doubell T. The pathophysiology of chronic paindincreased sensitivity to low threshold A-beta fiber inputs. Curr Opin Neurobiol 1994;4: Yakhnitsa V, Linderoth B, Meyerson BA. Effects of spinal cord stimulation on dorsal horn neuronal activity in a rat model of mononeuropathy. Pain 1999;79: Meyerson BA, Ren B, Herregodts P, Linderoth B. Spinal cord stimulation in animal models of mononeuropathy: effects on the withdrawal response and the flexor reflex. Pain 1995; 61: Saadé NE, Tabet MS, Banna NR, Atweh SF, Jabbur SJ. Inhibition of nociceptive evoked activity in spinal neurons through a dorsal column-brainstem-spinal loop. Brain Res 1985;339: Wallin J, Fiska A, Tjolsen A, Linderoth B, Hole K. Spinal cord stimulation inhibits long-term potentiation of spinal wide dynamic range neurons. Brain Res 2003;973: El-Khoury C, Hawwa M, Baliki SF, et al. Attenuation of neuropathic pain by segmental and supraspinal activation of the dorsal column system in awake rats. Neuroscience 2002;112: Meyerson BA, Brodin E, Linderoth B. Possible neurohumoral mechanisms in CNS stimulation for pain suppression. Appl Neurophysiol 1985;48: Linderoth B, Gazelius B, Franck J, Brodin E. Dorsal column stimulation induces release of serotonin and substance P in the cat dorsal horn. Neurosurgery 1992;31: Cui JG, O Connor WT, Ungerstedt U, Linderoth B, Meyerson BA. Spinal cord stimulation attenuates augmented dorsal horn release of excitatory amino acids in mononeuropathy via a GABAergic mechanism. Pain 1997;73: Stiller C-O, Cui J-G, O Connor WT, et al. Release of GABA in the dorsal horn and suppression of tactile allodynia by spinal cord stimulation in mononeuropathic rats. Neurosurgery 1996;39: Cui JG, Linderoth B, Meyerson BA. Effects of spinal cord stimulation on touch-evoked allodynia involve GABAergic mechanisms. An experimental study in the mononeuropathic rat. Pain 1996;66: Cui JG, Sollevi A, Linderoth B, Meyerson BA. Adenosine receptor activation suppresses tactile hypersensitivity and potentiates effect of spinal cord in mononeuropathic rats. Neurosci Lett 1997;223: Lind G, Meyerson BA, Winter J, Linderoth B. Intrathecal baclofen as adjuvant therapy to enhance the effect of spinal cord stimulation in neuropathic pain. A pilot study. Eur J Pain 2004;8: Wallin J, Cui J-G, Yahknitsa V, et al. Gabapentin and pregabalin suppress tactile allodynia and potentiate spinal cord stimulation in a model of neuropathy. Eur J Pain 2002;6: Schechtmann G, Wallin J, Meyerson BA, Linderoth B. Intrathecal clonidine potentiates suppression of tactile hypersensitivity by spinal cord stimulation in a model of neuropathy. Anesth Analg 2004;99: Obata H, Li X, Eisenach JC. Alpha2-adrenoceptor activation by clonidine enhances stimulation-evoked acetylcholine release from spinal cord tissue after nerve ligation in rats. Anesthesiology 2005; 102: Mao J. Translational pain research: bridging the gap between basic and clinical research. Pain 2002; 97: Baron R, Levine JD, Fields HL. Causalgia and reflex sympathetic dystrophy: does the sympathetic nervous system contribute to the generation of pain? Muscle Nerve 1999;22: Kemler MA, De Vet HC, Barendse GA, Van Den Wildenberg FA, Van Kleef M. Effect of spinal cord stimulation in patients with chronic reflex sympathetic dystrophy: two years follow-up of the randomized controlled trial. Ann Neurol 2004;55: Kumar K, Nath RK, Toth C. Spinal cord stimulation is effective in the management of reflex sympathetic dystrophy. Neurosurgery 1997;40: Olsson GL, Meyerson B. Spinal cord stimulation as a treatment option in severe cases of complex regional pain syndrome. 6th International Symposium on Paediatric Pain, Sydney June 2003; Abstract 236, p Linderoth B, Gunasekera L, Meyerson BA. Effects of sympathectomy on skin and muscle microcirculation during dorsal column stimulation: animal studies. Neurosurgery 1991;29: Croom JE, Foreman RD, Chandler MJ, Barron KW. Cutaneous vasodilatation during dorsal column stimulation is mediated by dorsal roots and CGRP. Am J Physiol 1997;272:H950--H Tanaka S, Komori N, Barron KW, et al. Mechanisms of sustained cutaneous vasodilation induced by spinal cord stimulation. Auton Neurosci 2004; 114: Kemler MA, Barendse GA, van Kleef M, Egbrink MG. Pain relief in complex regional pain syndrome due to spinal cord stimulation does not depend on vasodilation. Anesthesiology 2000;92: