- Email: [email protected]
Effects of pulsed versus conventional radiofrequency current on rabbit dorsal root ganglion morphology
European Journal of Pain 9 (2005) 251–256 www.EuropeanJournalPain.com
Effects of pulsed versus conventional radiofrequency current on rabbit dorsal root ganglion morphology Serdar Erdine
a,*
, Aysen Yucel a, Ali Cimen b, Salih Aydin c, Aydin Sav d, Ayhan Bilir
e
a
Department of Algology, Istanbul Faculty of Medicine, Istanbul University, Capa Klinikleri, Cerrahi Monoblok, 34390 Istanbul, Turkey Department of Anesthesiology, Istanbul Faculty of Medicine, Istanbul University, Capa Klinikleri, Cerrahi Monoblok, 34390 Istanbul, Turkey c Department of Neurosurgery, Istanbul Faculty of Medicine, Istanbul University, Capa Klinikleri, Cerrahi Monoblok, 34390 Istanbul, Turkey d Marmara University Institute of Neurological Sciences, Istanbul, Turkey e Department of Histology and Embryology, Istanbul Faculty of Medicine, Istanbul University, Temel Bilimler, 34390 Istanbul, Turkey
b
Received 17 March 2004; accepted 7 July 2004 Available online 12 September 2004
Abstract Lesioning using radiofrequency (RF) current has been increasingly used in clinical practice for the treatment of pain syndromes. Although formation of heat causing ‘‘thermocoagulation’’ of the nervous tissues is thought to be responsible of the clinical outcome, a more recent modality of RF application named pulsed radiofrequency (PRF) delivers the RF current without producing destructive levels of heat. In our study, we compared the effects of conventional RF (CRF) and PRF on rabbit dorsal root ganglion (DRG) morphology, including also control and sham operated groups. The setting of the experiment and the RF parameters used were similar to those used in current clinical practice. The specimens were analyzed both with light microscopy and electron microscopy, two weeks after the procedure. At the light microscopic level, all groups had preserved the normal DRG morphology and no differences were observed between them. In the electron microscopic analysis there were no pathological findings in the control and sham operated groups. But the ganglion cells in the RF groups had enlarged endoplasmic reticulum cisterns and increased number of cytoplasmic vacuoles which were more evident in the CRF group. Some of the ganglion cells in the CRF group had mitochondrial degeneration, nuclear membrane disorders or loss of nuclear membrane and neurolemma integrity. The myelinated and unmyelinated nerve fibers were of normal morphology in all groups. Our results suggest that PRF application is less destructive of cellular morphology than CRF at clinically used ‘‘doses’’. Before making certain judgements, more experimental and clinical studies should be planned. 2004 European Federation of Chapters of the International Association for the Study of Pain. Published by Elsevier Ltd. All rights reserved. Keywords: Radiofrequency; Pulsed; Conventional; Dorsal root ganglion; Morphology; Ultrastructure
1. Introduction Lesioning using radiofrequency (RF) current has been used in clinical practice for the treatment of pain * Corresponding author. Tel.: +90 212 324 01 48; fax: +90 212 283 74 66. E-mail address: [email protected] (S. Erdine).
syndromes since it is first described by Rosomoff et al. (1965). While the first applications of RF was percutaneus cordotomy (Rosomoff et al., 1965) and trigeminal ganglion rhizotomy (Sweet and Wepsic, 1974), Uematsu described use of the technique for the dorsal root ganglion (DRG) lesioning in 1977 (Uematsu, 1977). Being a minimally invasive and effective method, RF applications incorporating technological advancements have
1090-3801/$30 2004 European Federation of Chapters of the International Association for the Study of Pain. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ejpain.2004.07.002
252
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
become very popular in pain management (Sluijter, 2001). In the conventional application of RF (CRF), heat is produced in the tissues surrounding the RF electrode tip (Sluijter and Van Kleef, 1998; Cosman et al., 1984), and until late nineties formation of heat was thought to be responsible for the clinical outcome. The first published material to question the presumed role of heat was SlappendelÕs study in which he concluded that there were no differences between the clinical outcomes of CRF applications with the RF electrode tip temperatures reaching 40 and 67 C (Slappendel et al., 1997). Soon after, Sluijter et al. stated that the application of CRF with a tip temperature of 42 C was ineffective for the pain relief, a result inconsistent with SlappendelÕs results (Sluijter et al., 1998). In the same paper he described another method for the application of RF with low electrode tip temperatures, in which he applied the RF current in a pulsed fashion and this new method was named pulsed radiofrequency (PRF). Sluijter et al. proposed that the effect of PRF is due to the electromagnetic field (EMF) produced during the application. The published studies on PRF applications, which are commonly small clinical series, suggest that the effect of this new technique in relieving pain is as good as CRF, although there are some studies with contrary results (Shah and Racz, 2004; Geurts et al., 2003; Van Zundert et al., 2003; Munglani, 1999). These studies also suggest that PRF is a non-destructive method (Cohen and Foster, 2003; Sluijter, 2000; Day, 1999; Munglani, 1999; Sluijter et al., 1998). However, the evidence for its efficacy and utility is very weak because it is documented mostly by non-controlled studies and, case reports (Shah and Racz, 2004; Van Zundert et al., 2003; Munglani, 1999). A placebo response to the pulsed RF procedure cannot be excluded in these studies and clearly these observations and case reports with pulsed RF are anecdotal and randomized controlled trials of this technique are required before any claims can be made about its place in the therapeutic armentarium. In the recent literature, there are two experimental studies investigating the tissue effects of RF as used for pain relief in patients (Higuchi et al., 2002; De Louw et al., 2001), but neither of them directly compared effects of CRF and PRF on neuron morphology at the ultra-structural level. Thus; there is no histological proof that PRF is a non-destructive method. In our study we compared the effects of CRF and PRF on rabbit DRG morphology, including also control and sham operated groups. The setting of the experiment and the RF parameters used were similar to those used in current clinical practice. The specimens were analyzed both with light microscopy and transmission electron microscopy.
2. Methods This study was conducted at Istanbul University, Istanbul Faculty of Medicine, Department of Algology, and Department of Histology and Embryology. Four New Zealand White Rabbits used in the study were supplied from Istanbul University Experimental Medical Research Institute and the study was carried out with the approval of the Animal Studies Ethics Committee of this institute. Following anesthesia with intramuscular (i.m.) 35 mg/kg ketamine and 5 mg/kg xylazine, the abdominal and lomber regions of the rabbits were shaved. The indifferent electrode of the RF device (Radionics 3C+, Radionics, Burlington, MA, USA) was attached to the abdominal skin of the rabbits. They were then positioned prone on the operation table, and the lumbar skin was prepared with antiseptic solution and draped sterilely. Following the confirmation of vertebral levels, the C-arm of the fluoroscope (Siemens, Siremobil, 2000, Germany) was rotated around its axis and made oblique such that the X-ray beam was parallel to the axis of the intervertebral foramen. With the C-arm in this position, a 5 cm, 22 G (0.7 mm) SMK cannula with 4 mm active tip (SMK-C5, Radionics, Burlington, MA, USA) was introduced transcutaneously until it rested at the dorsal part of the intervertebral foramen. After the placement was confirmed by anteroposterior (AP) and lateral fluoroscopic images (Fig. 1), the stylet of the cannula was replaced with an RF probe, tissue impedance was measured and the presence of muscle contractions was checked using 2 Hz electrical stimulation to a maximum of 2.0 V. If muscle contractions were observed with stimulation output lower than 1.0 V, the electrode was
Fig. 1. The placement of the electrode. Anteroposterior (AP) fluoroscopic image.
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
253
Table 1 Study groups and electron microscopy results Groups
Rabbits
Electron microscopic analysis
Group I (CRF)
First rabbit, right L1, L2, L3 DRGs Second rabbit, right L1, L2, L3 DRGs
Increased number of vacuoles Enlarged endoplasmic reticulum cisterns Integrity loss in the nuclear membrane Mitochondrial degeneration
Group II (PRF)
Third rabbit, left L1, L2, L3 DRGs Fourth rabbit, left L1, L2, L3 DRGs
Increased numbers of vacuoles Enlarged endoplasmic reticulum cisterns Normal nuclear membrane
Group III (Sham)
Second rabbit, left L1, L2, L3 DRGs Third rabbit, right L1, L2, L3 DRGs
No cellular pathology was observed
Group IV (Control)
First rabbit, left L1, L2, L3 DRGs Fourth rabbit, right L1, L2, L3 DRGs
No cellular pathology was observed
pulled back 1 mm. If muscle contractions were observed only with stimulation output higher than 1.3 V (Ford et al., 1984), or no contraction was observed, we rechecked fluoroscopic position of the electrode tip in AP, lateral and oblique positions, and then advanced the electrode another 1 mm. The procedure was repeated until muscle contractions were observed with stimulation output between 1.0 and 1.3 V. This criterion indicates that the electrode was near the DRG, but had not penetrated it. Following the proper electrode placement, the RF procedures were performed as noted below. A total of 6 DRGs in each rabbit, at segments L1, L2, L3; bilaterally were included in the study. Four experimental groups are described in Table 1. In Group I (CRF Group), a 60 s, 67 C lesion was made at each segmental level. In Group II (PRF Group) a 120 s PRF, at a rate of 2 Hz and an active cycle of 20 ms, was applied with 45 V generator output. During this procedure, if the tip temperature reading on the RF device exceeded 42 C, the output voltage was gradually decreased until the tip temperature fell below 43 C. In Group III (Sham Operation Group) no RF application was performed following the cannula and electrode placement. In Group IV (Control Group) no harware was inserted although fluoroscopy was performed. Two weeks after the RF applications, the rabbits were anesthetised with i.m. 35 mg/kg ketamine and 5 mg/kg xylazine. The thoracic and lumbar skin of the rabbits was shaved. Following thoracotomy, a 20 G cannula was inserted to the left ventricle and a 22 G cannula was inserted to the right atrium. For tissue fixation, 1% glutaraldehyde solution (at room temperature, in 0.1 M phosphate buffer, pH 7.4) was perfused from the left ventricle cannula until the fluid coming out the right atrial cannula was colorless. Immediately after glutaraldehyde perfusion, a laminectomy was carefully performed from T12 to L5, the spinal cord, DRGs and spinal nerves were dissected, and these structures were excised en bloc. The specimen was cut into pieces, each piece containing a DRG, and they were put into vials
containing 2.5% glutaraldehyde (in 0.1 M phosphate buffer, pH 7.4, T: 4 C). During every step of the dissection procedure, the tissues were irrigated with 1% gluteraldehyde solution (at room temperature, in 0.1 M phosphate buffer, pH 7.4). The gluteraldehyde solution used during the above mentioned fixation process was prepared just prior to the surgical procedures. The DRG specimens in 2.5% glutaraldehyde solution were kept at +4 C for 3 h and then postfixed with 1% osmium tetroxide (in phosphate buffer, pH 7.4, T: 4 C). This was followed by en bloc staining with aqueous 1% uranyl acetate and dehydration with a graded ethanol series. The specimens were embedded in epon mixture (Epoxy Embedding Medium, Fluka, cat. No: 45345; DDSA, Merck, cat. No: 1.12147; MNA, Merck, cat. No: 1.12251; DMP, Merck, cat. No: 12388) and following 18 h polymerization at 60 C, sequential semithin sections beginning from the proximal side of the specimens were prepared on the ultramicrotome (Reichert Om U3, Germany). Those sections were stained with toluidin blue and observed with light microscopy to determine the localization of the DRGs. After this,
Fig. 2. Ganglion cell, Control Group. No cellular pathology was observed (N: nucleus, NM: nuclear membrane, original magnification: 13,200·).
254
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
DRG containing regions were trimmed and sectioned to ˚ , placed on nickel grids, stained with a thickness of 700 A 5% uranyl acetate for 30 min and lead citrate for 10 min. The samples were then viewed on the transmission electron microscope (Jeol, JEM-100C, Japan). For the light microscopic analysis the specimens were embedded in paraffin blocks and following the haematoxylin and eosin staining of 3 lm thick sections, the samples were observed with light microscopy (Zeiss, Germany). All the morphological data was collected and interpreted by a histologist and a pathologist who were blind to the study groups.
dynia on handling or brushing the skin of the L1–L3 dermatomes, and no autotomy. 3.1. Light microscopic analysis At the light microscopic level DRGs contained oval shaped ganglion cells surrounded by a single layer of satellite cells, longitudinally sectioned myelinated axons and unmyelinated axon bundles, and endoneural vasculature. There were no signs of pathological changes. No differences were observed between the groups. 3.2. Electron microscopic analysis
3. Results After the RF lesioning procedures, no obvious motor deficit such as limb paralysis, limbing or altered gait was observed in any of the rabbits during the follow-up period. Likewise, there were no obvious indications of allo-
Fig. 3. Ganglion cell, CRF Group. (a) The nucleus membrane has lost its integrity and normal morphology. Together with mitochondrial degeneration, these findings suggest cellular viability loss (N: nucleus, NM: nuclear membrane, M: degenerated mitochondria, *: integrity loss in the nuclear membrane, original magnification: 32,000·). (b) Invagination of the nuclear membrane. Also the numbers of vacuoles are increased (N: nucleus, NM: nuclear membrane, *: invaginated nuclear membrane, original magnification: 16,600·).
In Group III and Group IV DRGs, the ganglion cells, satellite cells, myelinated and unmyelinated nerve fibers had normal structure and localization. No pathological findings were observed in the cytoplasm, euchromatic nucleus, neurolemma and nuclear membrane of the ganglion cells (Fig. 2). Specifically endoplasmic reticulum cisterns were of normal morphology. There was no evident difference between these two groups.
Fig. 4. Ganglion cell, PRF Group. (a) The endoplasmic reticulum cisterns are enlarged and the numbers of vacuoles are increased (ER: endoplasmic reticulum, original magnification: 13,200·). (b) The nuclear membrane has its normal structure and integrity (N: nucleus, NM: nuclear membrane, original magnification: 16,600·).
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
In contrast, RF current had marked effects on DRG ultrastructure. DRG neurons in Group I contained numerous giant cytoplasmic vacuoles, fused with each other. It was also observed that the endoplasmic reticulum cisterns with evident ribosomes were extremely enlarged, the crista formations of the mitochondria are diminished and the mitochondria are degenerated in the ganglion cells. There were nuclear membrane disorders or loss of nuclear membrane and neurolemma integrity in some of the ganglion cells (Fig. 3(a) and (b)). No evident morphological pathology was observed in the myelinated and unmyelinated nerve fibers. In DRG neurons in Group II, enlargements in the endoplasmic reticulum cisterns of the ganglion cells were observed (Fig. 4(a)). The ganglion cells also revealed vacuole groups having the tendency of fusion with each other, localized linearly in the neuroplasma close to the neurolemma. There were no structural pathology in the cell membranes (Fig. 4(b)). The myelinated and unmyelinated nerve fibers were of normal morphology.
4. Discussion In this study, our main goal was to examine the effects of RF current on the DRG. Our results suggest that PRF application is much less destructive of cellular morphology than CRF at clinically used ‘‘doses’’. However DRGs subjected to PRF did show some ultrastructural changes in cell morphology. In previous studies, it was concluded that RF applications caused some structural damage on myelinated, unmyelinated nerve fibers or both. Letcher and Goldring (1968) suggested a fiber-selective effect of heat on the nervous tissue by mentioning small myelinated fiber destruction while Uematsu (1977) and Smith et al. (1981) suggested an indiscriminate destruction of both small and large fibers. Kanpolat and Onol (1980) have reported massive necrosis in ganglion cells 14 days after trigeminal RF which was performed in dog. In a recent study by De Louw et al. (2001), no myelinated and unmyelinated nerve damage was observed by CRF application. They concluded that damage observed by Letcher, Uematsu and Smith was due to the facts that the techniques performed were using electrodes much larger than are currently used in clinical practice and the electrodes were positioned inside the nervous tissue. De Louw et al. used modern, higher gauge electrodes and applied the RF current adjacent to the DRG, not through electrodes placed within the ganglion as in earlier studies. In their report De Louw et al. state that there was no observable damage in the DRG at the light microscopic level although they mention a significant increase in MIB-1 activity, which may be a marker of cellular damage.
255
We investigated the effects of both CRF and PRF on the DRGs at the light and electron microscopic levels. Our light microscopic results may be compared with those of De Louw et al. (2001), since it is the only published study using RF settings similar to ours, and using equipment and electrodes like those used clinically. We did not see any ganglion cell, myelinated or unmyelinated nerve fiber damage in the DRGs, in concordance with the De Louw et al.Õs study. In our study, although there were no significant differences between the groups under the light microscope, the electron microscopic analysis revealed some structural differences such as; giant cytoplasmic vacuoles, extremely enlarged endoplasmic reticulum cisterns, degenerated mitochondria in the ganglion cells and loss of nuclear membrane within the ganglion cells in CRF group. The differences in the CRF group, some of which may suggest cellular viability loss, may be a result of heat, higher EMF formation, or both. Prospective and randomized studies of RF with large sample sizes are lacking. Published clinical studies have claimed that PRF is non-destructive on the grounds that no sensory deficiency was observed with the application of PRF (Cohen, 2003; Sluijter, 2001, 2000; Sluijter et al., 1998). Our morphological analysis supports the clinical findings as we did not observe any visible cell loss attributable to the RF application in the PRF Group, and ultrastructural damage was modest. In conclusion, PRF application may be regarded substantially less destructive than CRF. Although the light microscopic analysis showed no obvious difference between these two groups, the electron microscopic analysis did reveal significant differences between the PRF and CRF groups. We still do not know much about the effects of EMF, produced by the RF current in the treatment of pain syndromes. These changes in the cellular components may be attributable to this effect and responsible for the clinical outcome. Before making certain judgements, more experimental and clinical studies should be planned.
References Cohen SP, Foster A. Pulsed radiofrequency as a treatment for groin pain and orchialgia. Urology 2003;61:45. Cosman ER, Nashold BS, Ovelmann-Levitt J. Theoretical aspects of radiofrequency lesions in the dorsal root entry zone. Neurosurgery 1984;15:945–50. Day M. Sphenopalatine ganglion analgesia. Curr Rev Pain 1999;3:342–7. De Louw AJA, Vles HSH, Freling G, Herpers MJHM, Van Kleef M. The morphological effects of a radiofrequency lesion adjacent to the dorsal root ganglion (RF-DRG) – an experimental study in the goat. Eur J Pain 2001;5:169–74. Ford DJ, Pither C, Raj PP. Comparison of insulated and uninsulated needles for locating peripheral nerves with a peripheral nerve stimulator. Anesth Analg 1984;63(10):925–8.
256
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
Geurts JW, van Wijk RM, Wynne HJ, Hammink E, Buskens E, Lousberg R, et al. Radiofrequency lesioning of dorsal root ganglia for chronic lumbosacral radicular pain: a randomised, doubleblind, controlled trial. Lancet 2003;361(9351):21–6. Higuchi Y, Nashold B, Sluijter M, Cosman E, Pearlstein RD. Exposure of the dorsal root ganglion in rats to pulsed radiofrequency currents activates dorsal horn lamina I and II neurons. Neurosurgery 2002;50:850–5. Kanpolat Y, Onol B. Experimental percutaneous approach to the trigeminal ganglion in dogs with histopathological evaluation of radiofrequency lesions. Acta Neurochir Suppl 1980;30: 363–6. Letcher FS, Goldring S. The effect of radiofrequency current and heat on peripheral nerve action potential in the cat. J Neurosurg 1968;29:42–7. Munglani R. The longer term effect of pulsed radiofrequency for neuropathic pain. Pain 1999;80:437–9. Rosomoff HL, Carroll F, Brown J, Sheptak P. Percutaneous radiofrequency cervical cordotomy technique. J Neurosurg 1965;23: 639–44. Shah RV, Racz GB. Long-term relief of posttraumatic headache by sphenopalatine ganglion pulsed radiofrequency lesioning: a case report. Arch Phys Med Rehabil 2004;85(6):1013–6. Slappendel R, Crul BJP, Braak GJJ, Geurts JWM, Booij LHDJ, Voerman VF, et al. The efficacy of radiofrequency lesioning of the
cervical spinal dorsal root ganglion in a double blinded randomized study: no difference between 40 and 67 C treatments. Pain 1997;73:159–63. Sluijter ME. The role of radiofrequency in failed back surgery patients. Curr Rev Pain 2000;4:49–53. Sluijter ME. Radiofrequency; Part 1: The lumbosacral region. Switzerland, SA: FlivoPress; 2001, p. 49–72. Sluijter ME, Van Kleef M. Characteristics and mode of action of radiofrequency lesions. Curr Rev Pain 1998;2:143–50. Sluijter ME, Cosman ER, Rittman WB, Van Kleef M. The effects of pulsed radiofrequency fields applied to the dorsal root ganglion – a preliminary report. Pain Clinic 1998;11:109–17. Smith HP, McWhorter JM, Challa VR. Radiofrequency neurolysis in a clinical model. J Neurosurg 1981;55:246–53. Sweet WH, Wepsic JG. Controlled thermocoagulation of trigeminal ganglion and rootlets for differential destruction of pain fibres. J Neurosurg 1974;40:143–56. Uematsu S. Percutaneus electrothermocoagulation of spinal nerve trunk, ganglion and rootlets. In: Schmidel HH, Sweet WS, editors. Current technique in operative neurosurgery. New York: Grune and Stratton; 1977. p. 469–90. Van Zundert J, Brabant S, Van de Kelft E, Vercruyssen A, Van Buyten JP. Pulsed radiofrequency treatment of the Gasserian ganglion in patients with idiopathic trigeminal neuralgia. Pain 2003;104(3): 449–52.
Effects of pulsed versus conventional radiofrequency current on rabbit dorsal root ganglion morphology Serdar Erdine
a,*
, Aysen Yucel a, Ali Cimen b, Salih Aydin c, Aydin Sav d, Ayhan Bilir
e
a
Department of Algology, Istanbul Faculty of Medicine, Istanbul University, Capa Klinikleri, Cerrahi Monoblok, 34390 Istanbul, Turkey Department of Anesthesiology, Istanbul Faculty of Medicine, Istanbul University, Capa Klinikleri, Cerrahi Monoblok, 34390 Istanbul, Turkey c Department of Neurosurgery, Istanbul Faculty of Medicine, Istanbul University, Capa Klinikleri, Cerrahi Monoblok, 34390 Istanbul, Turkey d Marmara University Institute of Neurological Sciences, Istanbul, Turkey e Department of Histology and Embryology, Istanbul Faculty of Medicine, Istanbul University, Temel Bilimler, 34390 Istanbul, Turkey
b
Received 17 March 2004; accepted 7 July 2004 Available online 12 September 2004
Abstract Lesioning using radiofrequency (RF) current has been increasingly used in clinical practice for the treatment of pain syndromes. Although formation of heat causing ‘‘thermocoagulation’’ of the nervous tissues is thought to be responsible of the clinical outcome, a more recent modality of RF application named pulsed radiofrequency (PRF) delivers the RF current without producing destructive levels of heat. In our study, we compared the effects of conventional RF (CRF) and PRF on rabbit dorsal root ganglion (DRG) morphology, including also control and sham operated groups. The setting of the experiment and the RF parameters used were similar to those used in current clinical practice. The specimens were analyzed both with light microscopy and electron microscopy, two weeks after the procedure. At the light microscopic level, all groups had preserved the normal DRG morphology and no differences were observed between them. In the electron microscopic analysis there were no pathological findings in the control and sham operated groups. But the ganglion cells in the RF groups had enlarged endoplasmic reticulum cisterns and increased number of cytoplasmic vacuoles which were more evident in the CRF group. Some of the ganglion cells in the CRF group had mitochondrial degeneration, nuclear membrane disorders or loss of nuclear membrane and neurolemma integrity. The myelinated and unmyelinated nerve fibers were of normal morphology in all groups. Our results suggest that PRF application is less destructive of cellular morphology than CRF at clinically used ‘‘doses’’. Before making certain judgements, more experimental and clinical studies should be planned. 2004 European Federation of Chapters of the International Association for the Study of Pain. Published by Elsevier Ltd. All rights reserved. Keywords: Radiofrequency; Pulsed; Conventional; Dorsal root ganglion; Morphology; Ultrastructure
1. Introduction Lesioning using radiofrequency (RF) current has been used in clinical practice for the treatment of pain * Corresponding author. Tel.: +90 212 324 01 48; fax: +90 212 283 74 66. E-mail address: [email protected] (S. Erdine).
syndromes since it is first described by Rosomoff et al. (1965). While the first applications of RF was percutaneus cordotomy (Rosomoff et al., 1965) and trigeminal ganglion rhizotomy (Sweet and Wepsic, 1974), Uematsu described use of the technique for the dorsal root ganglion (DRG) lesioning in 1977 (Uematsu, 1977). Being a minimally invasive and effective method, RF applications incorporating technological advancements have
1090-3801/$30 2004 European Federation of Chapters of the International Association for the Study of Pain. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ejpain.2004.07.002
252
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
become very popular in pain management (Sluijter, 2001). In the conventional application of RF (CRF), heat is produced in the tissues surrounding the RF electrode tip (Sluijter and Van Kleef, 1998; Cosman et al., 1984), and until late nineties formation of heat was thought to be responsible for the clinical outcome. The first published material to question the presumed role of heat was SlappendelÕs study in which he concluded that there were no differences between the clinical outcomes of CRF applications with the RF electrode tip temperatures reaching 40 and 67 C (Slappendel et al., 1997). Soon after, Sluijter et al. stated that the application of CRF with a tip temperature of 42 C was ineffective for the pain relief, a result inconsistent with SlappendelÕs results (Sluijter et al., 1998). In the same paper he described another method for the application of RF with low electrode tip temperatures, in which he applied the RF current in a pulsed fashion and this new method was named pulsed radiofrequency (PRF). Sluijter et al. proposed that the effect of PRF is due to the electromagnetic field (EMF) produced during the application. The published studies on PRF applications, which are commonly small clinical series, suggest that the effect of this new technique in relieving pain is as good as CRF, although there are some studies with contrary results (Shah and Racz, 2004; Geurts et al., 2003; Van Zundert et al., 2003; Munglani, 1999). These studies also suggest that PRF is a non-destructive method (Cohen and Foster, 2003; Sluijter, 2000; Day, 1999; Munglani, 1999; Sluijter et al., 1998). However, the evidence for its efficacy and utility is very weak because it is documented mostly by non-controlled studies and, case reports (Shah and Racz, 2004; Van Zundert et al., 2003; Munglani, 1999). A placebo response to the pulsed RF procedure cannot be excluded in these studies and clearly these observations and case reports with pulsed RF are anecdotal and randomized controlled trials of this technique are required before any claims can be made about its place in the therapeutic armentarium. In the recent literature, there are two experimental studies investigating the tissue effects of RF as used for pain relief in patients (Higuchi et al., 2002; De Louw et al., 2001), but neither of them directly compared effects of CRF and PRF on neuron morphology at the ultra-structural level. Thus; there is no histological proof that PRF is a non-destructive method. In our study we compared the effects of CRF and PRF on rabbit DRG morphology, including also control and sham operated groups. The setting of the experiment and the RF parameters used were similar to those used in current clinical practice. The specimens were analyzed both with light microscopy and transmission electron microscopy.
2. Methods This study was conducted at Istanbul University, Istanbul Faculty of Medicine, Department of Algology, and Department of Histology and Embryology. Four New Zealand White Rabbits used in the study were supplied from Istanbul University Experimental Medical Research Institute and the study was carried out with the approval of the Animal Studies Ethics Committee of this institute. Following anesthesia with intramuscular (i.m.) 35 mg/kg ketamine and 5 mg/kg xylazine, the abdominal and lomber regions of the rabbits were shaved. The indifferent electrode of the RF device (Radionics 3C+, Radionics, Burlington, MA, USA) was attached to the abdominal skin of the rabbits. They were then positioned prone on the operation table, and the lumbar skin was prepared with antiseptic solution and draped sterilely. Following the confirmation of vertebral levels, the C-arm of the fluoroscope (Siemens, Siremobil, 2000, Germany) was rotated around its axis and made oblique such that the X-ray beam was parallel to the axis of the intervertebral foramen. With the C-arm in this position, a 5 cm, 22 G (0.7 mm) SMK cannula with 4 mm active tip (SMK-C5, Radionics, Burlington, MA, USA) was introduced transcutaneously until it rested at the dorsal part of the intervertebral foramen. After the placement was confirmed by anteroposterior (AP) and lateral fluoroscopic images (Fig. 1), the stylet of the cannula was replaced with an RF probe, tissue impedance was measured and the presence of muscle contractions was checked using 2 Hz electrical stimulation to a maximum of 2.0 V. If muscle contractions were observed with stimulation output lower than 1.0 V, the electrode was
Fig. 1. The placement of the electrode. Anteroposterior (AP) fluoroscopic image.
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
253
Table 1 Study groups and electron microscopy results Groups
Rabbits
Electron microscopic analysis
Group I (CRF)
First rabbit, right L1, L2, L3 DRGs Second rabbit, right L1, L2, L3 DRGs
Increased number of vacuoles Enlarged endoplasmic reticulum cisterns Integrity loss in the nuclear membrane Mitochondrial degeneration
Group II (PRF)
Third rabbit, left L1, L2, L3 DRGs Fourth rabbit, left L1, L2, L3 DRGs
Increased numbers of vacuoles Enlarged endoplasmic reticulum cisterns Normal nuclear membrane
Group III (Sham)
Second rabbit, left L1, L2, L3 DRGs Third rabbit, right L1, L2, L3 DRGs
No cellular pathology was observed
Group IV (Control)
First rabbit, left L1, L2, L3 DRGs Fourth rabbit, right L1, L2, L3 DRGs
No cellular pathology was observed
pulled back 1 mm. If muscle contractions were observed only with stimulation output higher than 1.3 V (Ford et al., 1984), or no contraction was observed, we rechecked fluoroscopic position of the electrode tip in AP, lateral and oblique positions, and then advanced the electrode another 1 mm. The procedure was repeated until muscle contractions were observed with stimulation output between 1.0 and 1.3 V. This criterion indicates that the electrode was near the DRG, but had not penetrated it. Following the proper electrode placement, the RF procedures were performed as noted below. A total of 6 DRGs in each rabbit, at segments L1, L2, L3; bilaterally were included in the study. Four experimental groups are described in Table 1. In Group I (CRF Group), a 60 s, 67 C lesion was made at each segmental level. In Group II (PRF Group) a 120 s PRF, at a rate of 2 Hz and an active cycle of 20 ms, was applied with 45 V generator output. During this procedure, if the tip temperature reading on the RF device exceeded 42 C, the output voltage was gradually decreased until the tip temperature fell below 43 C. In Group III (Sham Operation Group) no RF application was performed following the cannula and electrode placement. In Group IV (Control Group) no harware was inserted although fluoroscopy was performed. Two weeks after the RF applications, the rabbits were anesthetised with i.m. 35 mg/kg ketamine and 5 mg/kg xylazine. The thoracic and lumbar skin of the rabbits was shaved. Following thoracotomy, a 20 G cannula was inserted to the left ventricle and a 22 G cannula was inserted to the right atrium. For tissue fixation, 1% glutaraldehyde solution (at room temperature, in 0.1 M phosphate buffer, pH 7.4) was perfused from the left ventricle cannula until the fluid coming out the right atrial cannula was colorless. Immediately after glutaraldehyde perfusion, a laminectomy was carefully performed from T12 to L5, the spinal cord, DRGs and spinal nerves were dissected, and these structures were excised en bloc. The specimen was cut into pieces, each piece containing a DRG, and they were put into vials
containing 2.5% glutaraldehyde (in 0.1 M phosphate buffer, pH 7.4, T: 4 C). During every step of the dissection procedure, the tissues were irrigated with 1% gluteraldehyde solution (at room temperature, in 0.1 M phosphate buffer, pH 7.4). The gluteraldehyde solution used during the above mentioned fixation process was prepared just prior to the surgical procedures. The DRG specimens in 2.5% glutaraldehyde solution were kept at +4 C for 3 h and then postfixed with 1% osmium tetroxide (in phosphate buffer, pH 7.4, T: 4 C). This was followed by en bloc staining with aqueous 1% uranyl acetate and dehydration with a graded ethanol series. The specimens were embedded in epon mixture (Epoxy Embedding Medium, Fluka, cat. No: 45345; DDSA, Merck, cat. No: 1.12147; MNA, Merck, cat. No: 1.12251; DMP, Merck, cat. No: 12388) and following 18 h polymerization at 60 C, sequential semithin sections beginning from the proximal side of the specimens were prepared on the ultramicrotome (Reichert Om U3, Germany). Those sections were stained with toluidin blue and observed with light microscopy to determine the localization of the DRGs. After this,
Fig. 2. Ganglion cell, Control Group. No cellular pathology was observed (N: nucleus, NM: nuclear membrane, original magnification: 13,200·).
254
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
DRG containing regions were trimmed and sectioned to ˚ , placed on nickel grids, stained with a thickness of 700 A 5% uranyl acetate for 30 min and lead citrate for 10 min. The samples were then viewed on the transmission electron microscope (Jeol, JEM-100C, Japan). For the light microscopic analysis the specimens were embedded in paraffin blocks and following the haematoxylin and eosin staining of 3 lm thick sections, the samples were observed with light microscopy (Zeiss, Germany). All the morphological data was collected and interpreted by a histologist and a pathologist who were blind to the study groups.
dynia on handling or brushing the skin of the L1–L3 dermatomes, and no autotomy. 3.1. Light microscopic analysis At the light microscopic level DRGs contained oval shaped ganglion cells surrounded by a single layer of satellite cells, longitudinally sectioned myelinated axons and unmyelinated axon bundles, and endoneural vasculature. There were no signs of pathological changes. No differences were observed between the groups. 3.2. Electron microscopic analysis
3. Results After the RF lesioning procedures, no obvious motor deficit such as limb paralysis, limbing or altered gait was observed in any of the rabbits during the follow-up period. Likewise, there were no obvious indications of allo-
Fig. 3. Ganglion cell, CRF Group. (a) The nucleus membrane has lost its integrity and normal morphology. Together with mitochondrial degeneration, these findings suggest cellular viability loss (N: nucleus, NM: nuclear membrane, M: degenerated mitochondria, *: integrity loss in the nuclear membrane, original magnification: 32,000·). (b) Invagination of the nuclear membrane. Also the numbers of vacuoles are increased (N: nucleus, NM: nuclear membrane, *: invaginated nuclear membrane, original magnification: 16,600·).
In Group III and Group IV DRGs, the ganglion cells, satellite cells, myelinated and unmyelinated nerve fibers had normal structure and localization. No pathological findings were observed in the cytoplasm, euchromatic nucleus, neurolemma and nuclear membrane of the ganglion cells (Fig. 2). Specifically endoplasmic reticulum cisterns were of normal morphology. There was no evident difference between these two groups.
Fig. 4. Ganglion cell, PRF Group. (a) The endoplasmic reticulum cisterns are enlarged and the numbers of vacuoles are increased (ER: endoplasmic reticulum, original magnification: 13,200·). (b) The nuclear membrane has its normal structure and integrity (N: nucleus, NM: nuclear membrane, original magnification: 16,600·).
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
In contrast, RF current had marked effects on DRG ultrastructure. DRG neurons in Group I contained numerous giant cytoplasmic vacuoles, fused with each other. It was also observed that the endoplasmic reticulum cisterns with evident ribosomes were extremely enlarged, the crista formations of the mitochondria are diminished and the mitochondria are degenerated in the ganglion cells. There were nuclear membrane disorders or loss of nuclear membrane and neurolemma integrity in some of the ganglion cells (Fig. 3(a) and (b)). No evident morphological pathology was observed in the myelinated and unmyelinated nerve fibers. In DRG neurons in Group II, enlargements in the endoplasmic reticulum cisterns of the ganglion cells were observed (Fig. 4(a)). The ganglion cells also revealed vacuole groups having the tendency of fusion with each other, localized linearly in the neuroplasma close to the neurolemma. There were no structural pathology in the cell membranes (Fig. 4(b)). The myelinated and unmyelinated nerve fibers were of normal morphology.
4. Discussion In this study, our main goal was to examine the effects of RF current on the DRG. Our results suggest that PRF application is much less destructive of cellular morphology than CRF at clinically used ‘‘doses’’. However DRGs subjected to PRF did show some ultrastructural changes in cell morphology. In previous studies, it was concluded that RF applications caused some structural damage on myelinated, unmyelinated nerve fibers or both. Letcher and Goldring (1968) suggested a fiber-selective effect of heat on the nervous tissue by mentioning small myelinated fiber destruction while Uematsu (1977) and Smith et al. (1981) suggested an indiscriminate destruction of both small and large fibers. Kanpolat and Onol (1980) have reported massive necrosis in ganglion cells 14 days after trigeminal RF which was performed in dog. In a recent study by De Louw et al. (2001), no myelinated and unmyelinated nerve damage was observed by CRF application. They concluded that damage observed by Letcher, Uematsu and Smith was due to the facts that the techniques performed were using electrodes much larger than are currently used in clinical practice and the electrodes were positioned inside the nervous tissue. De Louw et al. used modern, higher gauge electrodes and applied the RF current adjacent to the DRG, not through electrodes placed within the ganglion as in earlier studies. In their report De Louw et al. state that there was no observable damage in the DRG at the light microscopic level although they mention a significant increase in MIB-1 activity, which may be a marker of cellular damage.
255
We investigated the effects of both CRF and PRF on the DRGs at the light and electron microscopic levels. Our light microscopic results may be compared with those of De Louw et al. (2001), since it is the only published study using RF settings similar to ours, and using equipment and electrodes like those used clinically. We did not see any ganglion cell, myelinated or unmyelinated nerve fiber damage in the DRGs, in concordance with the De Louw et al.Õs study. In our study, although there were no significant differences between the groups under the light microscope, the electron microscopic analysis revealed some structural differences such as; giant cytoplasmic vacuoles, extremely enlarged endoplasmic reticulum cisterns, degenerated mitochondria in the ganglion cells and loss of nuclear membrane within the ganglion cells in CRF group. The differences in the CRF group, some of which may suggest cellular viability loss, may be a result of heat, higher EMF formation, or both. Prospective and randomized studies of RF with large sample sizes are lacking. Published clinical studies have claimed that PRF is non-destructive on the grounds that no sensory deficiency was observed with the application of PRF (Cohen, 2003; Sluijter, 2001, 2000; Sluijter et al., 1998). Our morphological analysis supports the clinical findings as we did not observe any visible cell loss attributable to the RF application in the PRF Group, and ultrastructural damage was modest. In conclusion, PRF application may be regarded substantially less destructive than CRF. Although the light microscopic analysis showed no obvious difference between these two groups, the electron microscopic analysis did reveal significant differences between the PRF and CRF groups. We still do not know much about the effects of EMF, produced by the RF current in the treatment of pain syndromes. These changes in the cellular components may be attributable to this effect and responsible for the clinical outcome. Before making certain judgements, more experimental and clinical studies should be planned.
References Cohen SP, Foster A. Pulsed radiofrequency as a treatment for groin pain and orchialgia. Urology 2003;61:45. Cosman ER, Nashold BS, Ovelmann-Levitt J. Theoretical aspects of radiofrequency lesions in the dorsal root entry zone. Neurosurgery 1984;15:945–50. Day M. Sphenopalatine ganglion analgesia. Curr Rev Pain 1999;3:342–7. De Louw AJA, Vles HSH, Freling G, Herpers MJHM, Van Kleef M. The morphological effects of a radiofrequency lesion adjacent to the dorsal root ganglion (RF-DRG) – an experimental study in the goat. Eur J Pain 2001;5:169–74. Ford DJ, Pither C, Raj PP. Comparison of insulated and uninsulated needles for locating peripheral nerves with a peripheral nerve stimulator. Anesth Analg 1984;63(10):925–8.
256
S. Erdine et al. / European Journal of Pain 9 (2005) 251–256
Geurts JW, van Wijk RM, Wynne HJ, Hammink E, Buskens E, Lousberg R, et al. Radiofrequency lesioning of dorsal root ganglia for chronic lumbosacral radicular pain: a randomised, doubleblind, controlled trial. Lancet 2003;361(9351):21–6. Higuchi Y, Nashold B, Sluijter M, Cosman E, Pearlstein RD. Exposure of the dorsal root ganglion in rats to pulsed radiofrequency currents activates dorsal horn lamina I and II neurons. Neurosurgery 2002;50:850–5. Kanpolat Y, Onol B. Experimental percutaneous approach to the trigeminal ganglion in dogs with histopathological evaluation of radiofrequency lesions. Acta Neurochir Suppl 1980;30: 363–6. Letcher FS, Goldring S. The effect of radiofrequency current and heat on peripheral nerve action potential in the cat. J Neurosurg 1968;29:42–7. Munglani R. The longer term effect of pulsed radiofrequency for neuropathic pain. Pain 1999;80:437–9. Rosomoff HL, Carroll F, Brown J, Sheptak P. Percutaneous radiofrequency cervical cordotomy technique. J Neurosurg 1965;23: 639–44. Shah RV, Racz GB. Long-term relief of posttraumatic headache by sphenopalatine ganglion pulsed radiofrequency lesioning: a case report. Arch Phys Med Rehabil 2004;85(6):1013–6. Slappendel R, Crul BJP, Braak GJJ, Geurts JWM, Booij LHDJ, Voerman VF, et al. The efficacy of radiofrequency lesioning of the
cervical spinal dorsal root ganglion in a double blinded randomized study: no difference between 40 and 67 C treatments. Pain 1997;73:159–63. Sluijter ME. The role of radiofrequency in failed back surgery patients. Curr Rev Pain 2000;4:49–53. Sluijter ME. Radiofrequency; Part 1: The lumbosacral region. Switzerland, SA: FlivoPress; 2001, p. 49–72. Sluijter ME, Van Kleef M. Characteristics and mode of action of radiofrequency lesions. Curr Rev Pain 1998;2:143–50. Sluijter ME, Cosman ER, Rittman WB, Van Kleef M. The effects of pulsed radiofrequency fields applied to the dorsal root ganglion – a preliminary report. Pain Clinic 1998;11:109–17. Smith HP, McWhorter JM, Challa VR. Radiofrequency neurolysis in a clinical model. J Neurosurg 1981;55:246–53. Sweet WH, Wepsic JG. Controlled thermocoagulation of trigeminal ganglion and rootlets for differential destruction of pain fibres. J Neurosurg 1974;40:143–56. Uematsu S. Percutaneus electrothermocoagulation of spinal nerve trunk, ganglion and rootlets. In: Schmidel HH, Sweet WS, editors. Current technique in operative neurosurgery. New York: Grune and Stratton; 1977. p. 469–90. Van Zundert J, Brabant S, Van de Kelft E, Vercruyssen A, Van Buyten JP. Pulsed radiofrequency treatment of the Gasserian ganglion in patients with idiopathic trigeminal neuralgia. Pain 2003;104(3): 449–52.