J Neurosurg 74:636-642, I991
Histological assessment of nerve lesions caused by epineurial electrode application in rat sciatic nerve WEANER GIRSCH, M.D., RUPERT KOLLER, M.D., HELMUT GRUBER, M.D., JURGEN HOLLE, M.D., CHRISTIAN LIEGL, UDO LOSERT, M.D.V., WINFRIED MAYa, D.PH., AND HERWIG THOMA, D.PH.
Second Surgical Clinic and Third Department ()f Anatomy, University of Vienna, Vienna, Austria u- The left sciatic nerve of 36 rats was exposed and four ring-shaped stainless steel wire electrodes were sutured to the epineurium of each nerve in the same manner as performed clinically for "carousel stimulation" in man. The rats were sacrificed 10 days (Group 1), 3 weeks (Group 2), or 3 months (Group 3) after implantation. The electrodes were excised, the nerves were embedded in Epon, and semithin sections were obtained for histological and planimetric assessment of lesions caused by the epineurially sutured electrodes. The right sciatic nerves served as controls. The total area of neural tissue within the perineurium was determined at three levels: at the site of the electrodes, 8 mm proximal, and 8 mm distal. The area of neural tissue damaged by the surgical procedure was expressed as a percentage of the total area. In Group 1, nine of 12 nerves showed lesions ranging from 0.39% to 25.39% of the total area of neural tissue, in Group 2 eight of 11 sciatic nerves showed lesions ranging from 0.24% to 13.03% of the total area, and in Group 3 five of 12 nerves showed lesions ranging from 0.21% to 4.96% of the total area. The pathologically altered areas in Groups 2 and 3 exhibited distinct signs of nerve fiber regeneration. The reasons for the decrease in damage from Group 1 to Group 3 and the clinical implications of the results for long-term electrical stimulation are discussed.
KEY WORDS morphometry 9 electrode, epineurial diaphragm pacing functional electrical stimulation "
F
O~ many years, functional electrical stimulation of peripheral nerves has been applied successfully in the restoration of limb movement or respiratory function in para- and tetraplegic patients. 4-7' 20-22 Since the first attempts at experimental and clinical use of this technique, many authors have investigated the question of whether functional electrical stimulation causes any change in the stimulated nerves. The effects of cuff electrodes, coiled wires, and intraneurally implanted electrodes on nerve integrity have been explored in many histological studies] -3'9"11.13.14,16.21-23 The so-called "carousel stimulation" was developed in 1972 with the aim of reducing fatigue of the stimulated muscles, m,2oIt requires the application of at least four electrodes as close to the nerve as possible. Annular stainless steel electrodes are sutured to the epineurium of the peripheral nerve in a circular manner using microsurgical techniques. Despite successful clinical application for the purpose of the "Vienna phrenic pacemaker, "2~ epineufially fixed ring electrodes have not become very common in functional electrical stimula636
.
nerve injury
9
tion compared to other methods such as cuffelectrodes. Previous experimental studies have revealed good functional and even good morphological results but have not provided satisfactory quantitative data concerning nerve alterations related to epineurial electrode applicarlon.8'~s The present study was undertaken in order to quantify the extent and time course of lesions caused by electrode application. In a statistically relevant number of rats, annular stainless steel electrodes were sutured to the epineurium of the sciatic nerve and the extent of lesions was determined by computer-assisted image analysis.
Materials and Methods Thirty-six female Sprague-Dawley rats, weighing an average of 255.7 gm each, were anesthetized with 12 to 14 mg ketamine hydrochloride/100 gm body weight for a standardized operation. Groups of four animals were housed together in cages and given a standard diet and water ad libilum.
J. Neurosurg. / Volume 74/April, 1991
Nerve lesions caused by epineural electrodes
FJf. 2. Schematic drawing showing arrangement of electrodes around the sciatic nerve and the levels of histological assessment used in this study. FIG. 1. Operative site at the time of electrode explantation in a Group 3 rat (3 months after placement). The electrodes are surrounded by tender connective tissue.
Electrode Implantation Ring-shaped electrodes, 1 mm in inner diameter and connected to an electrode lead 1 cm in length, were used. Electrode and lead were made from stainless steel; the lead was covered with a Silastic tube. The left sciatic nerve was exposed in all animals. With microsurgical techniques, the nerve was isolated from the surrounding tissue and the electrode was implanted. Four electrodes were positioned around the sciatic nerve within a distance of 5 m m from each other. The electrodes were fixed to the epineurium of the nerve by a 8-0 nylon suture (Fig. 1). The most proximal electrode was situated about 5 m m distal to the ischial tuberosity and the most distal electrode at a minimum distance of 3.5 m m proximal to the division of the sciatic nerve into its main branches (Fig. 2). The electrode leads were placed at a fight angle to the nerve between the dorsal and the adductor muscles of the thigh.
Experimental Groups The 36 animals were separated into three investigation groups consisting of 12 animals each. All 12 animals in each group received electrode implantation in the left sciatic nerve. For control data, two rats received microsurgical exposure of the right sciatic nerve (sham operation), and the right sciatic nerve was crushed in two other animals. Thus, the right sciatic nerve was exposed in 12 of the 36 animals. In six of these animals, the nerve was isolated from the surrounding tissue under the operating microscope, but electrode implantation was not carried out (sham operation); in six other rats the right sciatic nerve was crushed between the branches of a microsurgical forceps, and the remaining 24 animals received no surgical procedure on the fight sciatic nerve.
J. Neurosurg. / Volume 74/April, 1991
Explantation and Histological Examination The animals were sacrificed l0 days (Group 1), 3 weeks (Group 2), or 3 months (Group 3) after implantation. In each animal both sciatic nerves were exposed and excised between the sacrum and the knee joint, Under microscopic magnification, the four electrodes were removed from the epineurium of the left sciatic nerve. From each of the left sciatic nerves, specimens situated directly at the site of the electrodes and 8 m m proximal and 8 m m distal to them (Fig. 2) were taken for histological examination. The nerves of the right thigh served as controls and were harvested for histology at corresponding levels. The nerves were fixed in 3% glutaraldehyde, postfixed in 2% buffered osmium tetroxide, and embedded in Epon. For quantitative evaluation, 2-urn semithin cross sections were cut on an ultramicrotome and transmitted via television camera from the microscope to a personal computer for computer-assisted planimetry. The final magnification used for image analysis was x 780. Cross sections of the sciatic nerve at the three levels mentioned above were examined in regard to either signs of degeneration (such as myelin fragmentation, reduction of nerve fiber density, and increase in connective tissue) or signs of regeneration (for example, groups of small fibers with thinned myelin sheaths). The following parameters were measured for each specimen of the sciatic nerve: total cross-sectional area of neural tissue within the perineurial sheath; area of nerve segments showing signs of degeneration or regeneration; cross-sectional area of 500 normal-appearing nerve fibers (expressed as the calculated diameter of a circle of the same area); and cross-sectional area of the remaining intact and regenerating fibers in pathologically altered segments (expressed as the calculated diameter of a circle of the same area). The following data were calculated from these measurements: altered area proportional to total area of neural tissue within the perineurium (percentage of altered area), and neural cross-sectional area occupied by myelinated fibers pro637
W. Girsch, et al. TABLE 1 Number of sciatic nerves showing altered sectors at three time intervalsfollowing electrode implantation Factor
Group 1 Group 2 Group3 (10days) (3 wks) (3 mos) no. of rats 12 11 12 no. of It sciatic nerves with lesions 9 8 5 location of lesions proximal to electrode level 0 1 0 at electrodelevel 7 4 3 distal to electrode level 9 6 3
TABLE 2 Percentages of totat fascicular cross-sectional areas altered by electrode implantation* Locationof Groupl Group 2 Group 3 Altered Areas (10 days) (3 weeks) (3 months) at electrode 3.46%-12% 0.7%-13.03% 0.38%-4.96% site (6.65%) (5.15%) (1.6l %) distal to elec- 0.39%-25.4% 0.24%-5.74% 0.21%-2.36% trodes (6.00%) (4.49%) (0.65%) * Percentage expressed as range; numbers in parentheses denote the median.
Histological Findings in Experimental Nerves FIG. 3. Photomicrograph of a semithin cross section of a Group 1 rat sciatic nerve 8 mm distal to electrode application, 10 days after electrode implantation. No signs of nerve fiber degeneration can be seen. Calibration bar = 250 urn.
portional to that captured by connective tissue in normal and injured segments (nerve fiber density). The density and diameters of nerve fibers were evaluated to confirm the classification of normal-appearing or pathologically altered regions of the sciatic nerve. In cases where it was difficult to differentiate between intact sensory fascicles and areas in an advanced state of regeneration, small nerve fiber diameters in combination with remnants of degenerated myelin sheaths were the relevant criteria for classifying segments as pathologically altered.
Results Gross Findings At the time of autopsy, a distinct increase in connective tissue around the electrodes was found in six of the 35 rats independent of investigation group (one animal in Group 2 died 1 week after electrode implantation). In six rats one electrode and in one rat (Group 3) three electrodes had not preserved their original position in contact with the epineurium and were found more distant from the nerve. 638
Sciatic Nerve. At 8 m m proximal to the electrodes, the sciatic nerve was always composed of one fascicle with a mean cross-sectional area of 482,566 _+ 14,107 sq ~zm (standard error of the mean) within the perineurium. At the level of the electrodes two fascicles were usually found capturing a mean area of 505,533 + 10,818 sq urn. At 8 m m distal to the four electrodes, the sciatic nerve was composed of at least three fascicles due to the varying level of division into its muscular and cutaneous branches (Fig. 3). 17 The mean crosssectional area of the nerve at this site was 491,721 _+ 13,912 sq urn. N u m b e r o f Altered Nerves. The number of sciatic nerves showing pathologically altered segments was nine in Group 1, eight in Group 2, and five in Group 3 (Table 1). Lesions were seen at all levels of examination. Only one specimen showed an altered region proximal to the electrodes. At the level of the electrodes, 14 of the 35 nerves examined exhibited alterations. Distal to the site of the electrodes the number of sciatic nerves exhibiting pathologically altered segments was nine in Group 1, six in Group 2, and three in Group 3. In four cases (two in Group 1 and two in Group 2), alterations at the electrode level were not accompanied by altered nerve segments distal to the electrodes. Extent o f Lesions. At the site of the electrodes, between 3.46% and 12% of the total area was damaged by electrode implantation in Group 1 (Table 2). Pathologically altered areas ranged between 0.7% and 13.03% in Group 2, and between 0.38% and 4.96% in J. Neurosurg. / Volume 74/April, 1991
Nerve lesions caused by epineural electrodes TABLE 3 Mean percentages ~f total fascicular areas altered b)' electrode implantation* Location of Group 1 Group2 Group3 Altered Areas (10 days) (3 weeks) (3 months) at electrode site 4.74% 2.18% 0.57% distal to electrodes 6.43% 2.62% 0.27% * Altered areas include areas with and without detectable lesions.
Group 3. Distal to the site of electrodes the injured sectors occupied between 0.39% and 25.4% of the total area of neural tissue within the perineurium in Group 1. Corresponding ranges were 0.24% to 5.74% for Group 2 and 0.21% to 2.36% for Group 3. When all animals in each group are combined, the average pathologically altered areas at the electrode level involved 4.74% (Group 1), 2.18% (Group 2), and 0.57% (Group 3) of the total fascicular area (Table 3); corresponding values determined from cross sections distal to the electrodes were 6.43% in Group 1, 2.62% in Group 2, and 0.27% in Group 3. Appearance o f Lesions. Lesions referred to electrode application were usually confined to segments and never widely disseminated over the whole crosssectional area of the sciatic nerve. Normal and pathologically altered regions could be distinguished easily in Groups 1 and 2, whereas in Group 3 differentiation was not as apparent in some cases. In regions classified as normal (Fig. 3), the mean nerve fiber diameter was 6.91 _+ 0.09 um and the mean nerve fiber density (neural cross-sectional area occupied by myelinated fibers) was 64.57% + 1.02%. Regions classified as pathologically altered appeared differently in the various rat groups. Ten days after electrode implantation (Group 1), the injured sectors exhibited fragmented myelin sheaths, myelin globules, signs of myelin phagocytosis, and a reduction of nerve fiber density due to an increase in connective tissue and edematous swelling. Within these sectors a small amount of relatively large normal-appearing nerve fibers could be observed (Fig. 4A). NormaI and pathologically altered sectors were distinguished on the basis of the mean nerve fiber density, which was reduced to 10.97% _+ 1.44% in damaged areas. Three weeks after electrode application (Group 2), damaged nerve segments usually contained several small nerve fibers with thin myelin sheaths and remnants of degenerated myelin sheaths (Fig. 4B)) 8 In this group, the mean nerve fiber density was 16.71% + 1.22% and the mean nerve fiber diameter was 4.84 _ 0.46 um in pathologically altered areas. Both criteria allowed a clear distinction (Fig. 5). Three months after the implantation procedure (Group 3), remnants of degenerated myelin sheaths in combination with small nerve fibers indicated pathologically altered areas in an advanced state of regenerJ. Neurosurg. / Volume 74/April, 1991
FIc. 4. Photomicrograph of semithin cross sections of rat sciatic nerves 8 mm distal to electrode application demonstrating typical segmentally localized pathological alterations. Calibration bar = 50 urn. A: Ten days after electrode implantation, a segment of the common peroneal branch exhibits distinct signs of nerve fiber degeneration. The surrounding nerve fibers are intact. B: At 3 weeks after electrode implantation, the circumscribed area (in the center) contains several small regenerating nerve fibers. C: At 3 months after electrode implantation, a segment of a small fascicle exhibits relatively large regenerating nerve fibers. The adjacent tibial branch does not show any pathological alterations.
ation in four rats (Fig. 4C). However, in two specimens of this group the altered segments still exhibited pronounced fibrosis and only a few regenerating nerve fibers were seen. 639
W. Girsch, et al. Histological Findings in Control Specimens Normal Sciatic Nerves. In sciatic nerves on the right side, which had not received any surgical manipulation, the mean diameter was 6.83 _+ 0.17 um and the mean percentage of fascicular area occupied by nerve fibers was 62.7% _ 1.71%. Sham-Operated Nerves. No signs of degeneration or regeneration could be detected in the six right sciatic nerves that had been exposed but did not receive electrodes. The mean diameter of myelinated fibers and the mean percentage of nerve fiber density in these specimens were 7.81 + 0.19 um and 70.11% + 1.93%, respectively. Crush-Injured Nerves. Contrary. to nerve segments altered by electrode implantation, no intact nerve fibers were seen in sciatic nerves of Group 1 crushed 10 days before. Three weeks after being crushed, the regenerating sciatic nerves exhibited nerve fibers with a mean diameter of 3.37 _+ 0.13 um and a mean nerve fiber density percentage of 20.31% _+ 3.31%. The corresponding data 3 months after crushing was 5.44 _+ 0.19 um and 49.82% _+ 1.32%, respectively. In Groups 2 and 3 the appearance of alterations resulting from electrode application generally corresponded well with that of the specimens of the crushed right sciatic nerves in Group 1 (Fig. 5). Discussion The results of the present study reveal that ringshaped stainless steel electrodes sutured to the epineurium of a peripheral nerve do not alter the morpholog-
ical aspect of the nerve to a great extent, provided that microsurgical techniques are carefully used. The nerve lesions observed in this investigation were clearly visible, but in general were circumscribed, not extensive, and clearly decreased over time.
Previous Studies In all previous studies dealing with the influence of cuff, coiled wire, or intraneurally implanted electrodes on nerve integrity, nerve fiber damage and myelin degeneration, or at least an increase in connective tissue due to electrode application was observed. ~-3'8'9"j1-16.23 Thus, at least minimal morphological alterations of nervous tissue seem to be an inevitable feature of functional electrical stimulation in peripheral nerves, even if lesions do not always occur. In some of the previous studies 2'3'~'~3-~5 an electrical current was applied to the nerves prior to histological examination. The question remained whether the influence of the electrical current or mechanical factors were responsible for the alterations observed in peripheral nerves after functional electrical stimulation. According to relevant publications, alterations of nervous tissue are obviously dependent on the intensity and the frequency of electrical current. The present study was undertaken in order to quantify morphological lesions caused by epineurial electrode application. Since no electrical current has been applied, lesions can only be due to mechanical factors, either the operative procedure or chronic mechanical irritation by the implanted electrodes. Arrangement of Lesions It may be surprising on first sight that lesions associated with electrode application were usually arranged in sectors or segments and were never widely disseminated over the whole cross-sectional area of the nerve. According to relevant publications about nerve injuriess there is no intermingling of fibers in the distal portion of a nerve trunk. Our findings provide an excellent explanation as to why mechanical stress to a segment of the nerve results in segmentally localized lesions at a more distal level. In the some cases, lesions at the electrode level were not followed by detectable alterations distal to the site of the electrodes. Thus, there is every reason to believe that in these cases the mechanical stress of electrode implantation did not cause Wallerian degeneration but resulted in segmental demyelination.~9
FlG. 5. Frequency histogram showing diameters of 250 regenerating (open bars) and an equal number of normalappearing (filled bars) myelinated nerve fibers found in the left sciatic nerve at the electrode level of a Group 2 rat (3 weeks after electrode placement). 640
Time Course of Lesions Both the number and extent of damaged areas decreased distinctly from Group 1 to Group 3, possibly due to full nerve regeneration in areas that had been only slightly damaged by the implantation procedure. Since the Group 3 animals had undergone electrode implantation before the two other investigation groups, an increase in the surgeon's experience is not responsiJ. Neurosurg. / Volume 74/April, 1991
Nerve lesions caused by epineural electrodes ble for the decrease in the number and extent of lesions from Group 1 to Group 3. The persistence of fibrous tissue and reduction in nerve fiber density in two Group 3 specimens can be referred either to mechanical stress by the implanted electrodes or to heavy damage during the surgical procedure prohibiting proper regeneration in the 3-month period. ~9 Although the sciatic nerves exhibited distinctly visible lesions 10 days after electrode implantation, in only two cases did the extent of altered areas exceed 12% of the total area of the nerve. Three months after electrode implantation, none of the damaged nerve segments covered more than 5%. Combining all specimens distal to the electrodes in Group 3 shows that, on average, more than 99% of the total area of neural tissue was intact. In addition, nearly all pathologically altered nerve regions in Groups 2 and 3 exhibited signs of nerve fiber regeneration.
Cause of Lesions The number and extent of damaged areas in Group 1 and the distinct decrease of these two parameters in Groups 2 and 3 lead to the conclusion that the observed alterations occur at implantation. Electrodes left in place over time probably neither lead to new alterations nor prohibit nerve fiber regeneration. No signs of degeneration were seen in the control nerves that had been exposed without electrode application. The procedure of suturing electrodes to the epineufium of a peripheral nerve seems to be the most important factor responsible for morphological alterations. This mechanical stress, probably due to manipulations with microsurgical instruments, is obviously not comparable to the crush injury of nerves on the control side. The crush lesions in the control nerves were, especially in Group 1, more severe than those observed in the reference group with electrode application.
Conclusions
Although no functional measurements have been performed in the present study, the extent of lesions and their decrease with time make long-term impairment of peripheral nerve function by the application of epineurially sutured electrodes extremely improbable. Of course, rat nerves may regenerate faster than human nerves, but the general ability to regenerate is the same in rats as in humans. Thus, the results obtained from rat nerves in this study should be transferable to human nerves in which electrodes are applied for "carousel stimulation." At the time of autopsy, nine of 140 electrodes were found to have become dislocated. All the other electrodes were fixed by soft connective tissue in their original position in contact with the epineurium. They would easily be removed by use of microsurgical techniques. Thus, epineurial electrode application seems to
J. Neurosurg. / Volume 74/April, 1991
be a method of fixing electrodes in a stable position to the nerve as close as possible without causing extensive and permanent alterations in nervous tissue. References
1. Agnew WF, McCreery DB, Yuen TGH, et al: Histologic and physiologic evaluation of electrically stimulated peripheral nerve: considerations for the selection of parameters. Ann Biomed Eng 17:39-60, 1989 2. Bowman BR, Erickson RC II: Acute and chronic implantation of coiled wire intraneural electrodes during cyclical electrical stimulation. Ann Biomed Eng 13:75-93, 1985 3. Breederveld RS: Electrical stimulation of motor nerves: results of animal experiment. Life Snpp Syst 2:183-188, 1984 4. Fodstad H: The Swedish experience in phrenic nerve stimulation. PACE 10:246-251, 1987 5. Glenn WWL, Holcomb WG, Hogan J, et al: Diaphragm pacing by radiofrequency transmission in the treatment of chronic ventilatory insufficiency. J Thor Cardiovasc Surg 66:505-520, 1973 6. Glenn WWL, Holcomb WG, Hogan JF, et al: Long-term stimulation of the phrenic nerve for diaphragm pacing, in Hambrecht FT, Reswick JB (eds): Functional Electrical Stimulation. New York: Marcel Decker, 1977, pp 97-112 7. Glenn WWL, Holcomb WG, McLaughlin AJ, et al: Total ventilatory support in a quadriplegic patient with radiofrequency electrophrenic respiration. N Engl J Med 286: 513-516, 1972 8. Happak W, Koller R, Girsch W, et al: Histological examination after epineural electrode application, in Wallinga W, Boom HBK, de Vries J (eds): Eleetrophysiologieal Kinesiology. Amsterdam: Excerpta Medica, 1988, pp 75-78 9. Herschberg PI, Sohn D, Agrawal GP, et al: Histologic changes in continuous, long-term electrical stimulation of a peripheral nerve. IEEE Trans Biomed Eng 14: 109-114, 1967 10. Holle J, Moritz E, Thoma H: Die Karussellstimulation, eine neue Methode zur Elektrophrenischen Langzeitbeatmung. Wien Kiln Wochenschr 86:23-27, 1974 11. Hughes GB, Bottomy MB, Dickins JRE, et al: A comparative study of neuropathologic changes following pulsed and direct current stimulation of mouse sciatic nerve. Am J Otolaryngol 1:378-385, 1980 12. Hughes GB, Bottomy MB, Jackson CG, et al: Myelin and axon degeneration following direct current peripheal nerve stimulation. A prospective controlled experimental study. Otolaryngol Head Neck Surg 89:767-775, 1981 13. Kim JH, Manuelidis EE, Glenn WWL, et al: Diaphragm pacing. Histopathologic changes in the phrenic nerves following long-term electrical stimulation. J Thor Cardiovasc Surg 72:602-608, 1976 14. Kim JH, Manuelidis EE, Glenn WWL, et al: Light and electron microscopic studies of phrenic nerves after longterm electrical stimulation. J Neurosurg 58:84-91, 1983 15. Rosenkranz D, Fenzl G, Holle J, et al: Influence of longterm low direct current on rat ischiatic nerves. Appl Neurophysio149:42-52, 1986 16. Rutten WLC, van Wier HJ, Put JHM, et al: Sensitivity, selectivity and bioacceptance of an intraneural multielectrode stimulation device in silicon technology, in Wallinga W, Boom HBK, de Vries J (eds): Electrophysiologteal Kinesiology. Amsterdam: Excerpta Medica, 1988, pp 135-139 641
W. Girsch, et al. 17. Schmalbruch H: Fiber composition of the rat sciatic nerve. Anal Rec 215:71-81, 1986 18. SchrOder JM: Altered ratio between axon diameter and myelin sheath thickness in regenerated neF,'e fibers, Brain Res 45:49-65, 1972 19. Sunderland S: Nerves and Nerve Injuries, ed 2. New York: Churchill Livingstone, 1978, pp 31-132 20. Thoma H, Gerner H, Girsch W, et al: lmplantable neurostimulators: the phrenie pacemaker-technology and rehabilitation strategies, in Wallinga W, Boom HBK. de Vries J (eds): Electrophysiologieal Kinesiology. Amsterdam: Exeerpta Mediea, 1988, pp 143-152 21. Waters RL, McNeal D, Perry' J: Experimental correction of footdrop by electrical stimulation of the peroneal nerve. J Bone Joint Surg (Am) 57:1047-1054, 1975
642
22, Young RF: Diaphragm pacing asan adjunct in respiratory insufficiency. Neurosurgery 2:43-46, 1978 23. Yuen TGH, Agnew WE, Bullara LA: Histopathological evaluation of dog sacral nerve after chronic electrical stimulation for micturition. Neurosurgery 14:449-455, 1984 Manuscript received April 1[, 1990. Accepted in final form September 21, 1990. This work was supported by the "Lorenz-Bohler Foundation" and the Austrian Research Foundations, FWF and FFF, the Ministu of Research and Science, and the Austrian National Bank. Address reprinl requests to." Werner Girsch, M.D., 2rid Surgical University Clinic, Spitalgasse 23, A-1090 Vienna, Austria.
J. Neurosurg. / Volume 74/April, 1991
Histological assessment of nerve lesions caused by epineurial electrode application in rat sciatic nerve WEANER GIRSCH, M.D., RUPERT KOLLER, M.D., HELMUT GRUBER, M.D., JURGEN HOLLE, M.D., CHRISTIAN LIEGL, UDO LOSERT, M.D.V., WINFRIED MAYa, D.PH., AND HERWIG THOMA, D.PH.
Second Surgical Clinic and Third Department ()f Anatomy, University of Vienna, Vienna, Austria u- The left sciatic nerve of 36 rats was exposed and four ring-shaped stainless steel wire electrodes were sutured to the epineurium of each nerve in the same manner as performed clinically for "carousel stimulation" in man. The rats were sacrificed 10 days (Group 1), 3 weeks (Group 2), or 3 months (Group 3) after implantation. The electrodes were excised, the nerves were embedded in Epon, and semithin sections were obtained for histological and planimetric assessment of lesions caused by the epineurially sutured electrodes. The right sciatic nerves served as controls. The total area of neural tissue within the perineurium was determined at three levels: at the site of the electrodes, 8 mm proximal, and 8 mm distal. The area of neural tissue damaged by the surgical procedure was expressed as a percentage of the total area. In Group 1, nine of 12 nerves showed lesions ranging from 0.39% to 25.39% of the total area of neural tissue, in Group 2 eight of 11 sciatic nerves showed lesions ranging from 0.24% to 13.03% of the total area, and in Group 3 five of 12 nerves showed lesions ranging from 0.21% to 4.96% of the total area. The pathologically altered areas in Groups 2 and 3 exhibited distinct signs of nerve fiber regeneration. The reasons for the decrease in damage from Group 1 to Group 3 and the clinical implications of the results for long-term electrical stimulation are discussed.
KEY WORDS morphometry 9 electrode, epineurial diaphragm pacing functional electrical stimulation "
F
O~ many years, functional electrical stimulation of peripheral nerves has been applied successfully in the restoration of limb movement or respiratory function in para- and tetraplegic patients. 4-7' 20-22 Since the first attempts at experimental and clinical use of this technique, many authors have investigated the question of whether functional electrical stimulation causes any change in the stimulated nerves. The effects of cuff electrodes, coiled wires, and intraneurally implanted electrodes on nerve integrity have been explored in many histological studies] -3'9"11.13.14,16.21-23 The so-called "carousel stimulation" was developed in 1972 with the aim of reducing fatigue of the stimulated muscles, m,2oIt requires the application of at least four electrodes as close to the nerve as possible. Annular stainless steel electrodes are sutured to the epineurium of the peripheral nerve in a circular manner using microsurgical techniques. Despite successful clinical application for the purpose of the "Vienna phrenic pacemaker, "2~ epineufially fixed ring electrodes have not become very common in functional electrical stimula636
.
nerve injury
9
tion compared to other methods such as cuffelectrodes. Previous experimental studies have revealed good functional and even good morphological results but have not provided satisfactory quantitative data concerning nerve alterations related to epineurial electrode applicarlon.8'~s The present study was undertaken in order to quantify the extent and time course of lesions caused by electrode application. In a statistically relevant number of rats, annular stainless steel electrodes were sutured to the epineurium of the sciatic nerve and the extent of lesions was determined by computer-assisted image analysis.
Materials and Methods Thirty-six female Sprague-Dawley rats, weighing an average of 255.7 gm each, were anesthetized with 12 to 14 mg ketamine hydrochloride/100 gm body weight for a standardized operation. Groups of four animals were housed together in cages and given a standard diet and water ad libilum.
J. Neurosurg. / Volume 74/April, 1991
Nerve lesions caused by epineural electrodes
FJf. 2. Schematic drawing showing arrangement of electrodes around the sciatic nerve and the levels of histological assessment used in this study. FIG. 1. Operative site at the time of electrode explantation in a Group 3 rat (3 months after placement). The electrodes are surrounded by tender connective tissue.
Electrode Implantation Ring-shaped electrodes, 1 mm in inner diameter and connected to an electrode lead 1 cm in length, were used. Electrode and lead were made from stainless steel; the lead was covered with a Silastic tube. The left sciatic nerve was exposed in all animals. With microsurgical techniques, the nerve was isolated from the surrounding tissue and the electrode was implanted. Four electrodes were positioned around the sciatic nerve within a distance of 5 m m from each other. The electrodes were fixed to the epineurium of the nerve by a 8-0 nylon suture (Fig. 1). The most proximal electrode was situated about 5 m m distal to the ischial tuberosity and the most distal electrode at a minimum distance of 3.5 m m proximal to the division of the sciatic nerve into its main branches (Fig. 2). The electrode leads were placed at a fight angle to the nerve between the dorsal and the adductor muscles of the thigh.
Experimental Groups The 36 animals were separated into three investigation groups consisting of 12 animals each. All 12 animals in each group received electrode implantation in the left sciatic nerve. For control data, two rats received microsurgical exposure of the right sciatic nerve (sham operation), and the right sciatic nerve was crushed in two other animals. Thus, the right sciatic nerve was exposed in 12 of the 36 animals. In six of these animals, the nerve was isolated from the surrounding tissue under the operating microscope, but electrode implantation was not carried out (sham operation); in six other rats the right sciatic nerve was crushed between the branches of a microsurgical forceps, and the remaining 24 animals received no surgical procedure on the fight sciatic nerve.
J. Neurosurg. / Volume 74/April, 1991
Explantation and Histological Examination The animals were sacrificed l0 days (Group 1), 3 weeks (Group 2), or 3 months (Group 3) after implantation. In each animal both sciatic nerves were exposed and excised between the sacrum and the knee joint, Under microscopic magnification, the four electrodes were removed from the epineurium of the left sciatic nerve. From each of the left sciatic nerves, specimens situated directly at the site of the electrodes and 8 m m proximal and 8 m m distal to them (Fig. 2) were taken for histological examination. The nerves of the right thigh served as controls and were harvested for histology at corresponding levels. The nerves were fixed in 3% glutaraldehyde, postfixed in 2% buffered osmium tetroxide, and embedded in Epon. For quantitative evaluation, 2-urn semithin cross sections were cut on an ultramicrotome and transmitted via television camera from the microscope to a personal computer for computer-assisted planimetry. The final magnification used for image analysis was x 780. Cross sections of the sciatic nerve at the three levels mentioned above were examined in regard to either signs of degeneration (such as myelin fragmentation, reduction of nerve fiber density, and increase in connective tissue) or signs of regeneration (for example, groups of small fibers with thinned myelin sheaths). The following parameters were measured for each specimen of the sciatic nerve: total cross-sectional area of neural tissue within the perineurial sheath; area of nerve segments showing signs of degeneration or regeneration; cross-sectional area of 500 normal-appearing nerve fibers (expressed as the calculated diameter of a circle of the same area); and cross-sectional area of the remaining intact and regenerating fibers in pathologically altered segments (expressed as the calculated diameter of a circle of the same area). The following data were calculated from these measurements: altered area proportional to total area of neural tissue within the perineurium (percentage of altered area), and neural cross-sectional area occupied by myelinated fibers pro637
W. Girsch, et al. TABLE 1 Number of sciatic nerves showing altered sectors at three time intervalsfollowing electrode implantation Factor
Group 1 Group 2 Group3 (10days) (3 wks) (3 mos) no. of rats 12 11 12 no. of It sciatic nerves with lesions 9 8 5 location of lesions proximal to electrode level 0 1 0 at electrodelevel 7 4 3 distal to electrode level 9 6 3
TABLE 2 Percentages of totat fascicular cross-sectional areas altered by electrode implantation* Locationof Groupl Group 2 Group 3 Altered Areas (10 days) (3 weeks) (3 months) at electrode 3.46%-12% 0.7%-13.03% 0.38%-4.96% site (6.65%) (5.15%) (1.6l %) distal to elec- 0.39%-25.4% 0.24%-5.74% 0.21%-2.36% trodes (6.00%) (4.49%) (0.65%) * Percentage expressed as range; numbers in parentheses denote the median.
Histological Findings in Experimental Nerves FIG. 3. Photomicrograph of a semithin cross section of a Group 1 rat sciatic nerve 8 mm distal to electrode application, 10 days after electrode implantation. No signs of nerve fiber degeneration can be seen. Calibration bar = 250 urn.
portional to that captured by connective tissue in normal and injured segments (nerve fiber density). The density and diameters of nerve fibers were evaluated to confirm the classification of normal-appearing or pathologically altered regions of the sciatic nerve. In cases where it was difficult to differentiate between intact sensory fascicles and areas in an advanced state of regeneration, small nerve fiber diameters in combination with remnants of degenerated myelin sheaths were the relevant criteria for classifying segments as pathologically altered.
Results Gross Findings At the time of autopsy, a distinct increase in connective tissue around the electrodes was found in six of the 35 rats independent of investigation group (one animal in Group 2 died 1 week after electrode implantation). In six rats one electrode and in one rat (Group 3) three electrodes had not preserved their original position in contact with the epineurium and were found more distant from the nerve. 638
Sciatic Nerve. At 8 m m proximal to the electrodes, the sciatic nerve was always composed of one fascicle with a mean cross-sectional area of 482,566 _+ 14,107 sq ~zm (standard error of the mean) within the perineurium. At the level of the electrodes two fascicles were usually found capturing a mean area of 505,533 + 10,818 sq urn. At 8 m m distal to the four electrodes, the sciatic nerve was composed of at least three fascicles due to the varying level of division into its muscular and cutaneous branches (Fig. 3). 17 The mean crosssectional area of the nerve at this site was 491,721 _+ 13,912 sq urn. N u m b e r o f Altered Nerves. The number of sciatic nerves showing pathologically altered segments was nine in Group 1, eight in Group 2, and five in Group 3 (Table 1). Lesions were seen at all levels of examination. Only one specimen showed an altered region proximal to the electrodes. At the level of the electrodes, 14 of the 35 nerves examined exhibited alterations. Distal to the site of the electrodes the number of sciatic nerves exhibiting pathologically altered segments was nine in Group 1, six in Group 2, and three in Group 3. In four cases (two in Group 1 and two in Group 2), alterations at the electrode level were not accompanied by altered nerve segments distal to the electrodes. Extent o f Lesions. At the site of the electrodes, between 3.46% and 12% of the total area was damaged by electrode implantation in Group 1 (Table 2). Pathologically altered areas ranged between 0.7% and 13.03% in Group 2, and between 0.38% and 4.96% in J. Neurosurg. / Volume 74/April, 1991
Nerve lesions caused by epineural electrodes TABLE 3 Mean percentages ~f total fascicular areas altered b)' electrode implantation* Location of Group 1 Group2 Group3 Altered Areas (10 days) (3 weeks) (3 months) at electrode site 4.74% 2.18% 0.57% distal to electrodes 6.43% 2.62% 0.27% * Altered areas include areas with and without detectable lesions.
Group 3. Distal to the site of electrodes the injured sectors occupied between 0.39% and 25.4% of the total area of neural tissue within the perineurium in Group 1. Corresponding ranges were 0.24% to 5.74% for Group 2 and 0.21% to 2.36% for Group 3. When all animals in each group are combined, the average pathologically altered areas at the electrode level involved 4.74% (Group 1), 2.18% (Group 2), and 0.57% (Group 3) of the total fascicular area (Table 3); corresponding values determined from cross sections distal to the electrodes were 6.43% in Group 1, 2.62% in Group 2, and 0.27% in Group 3. Appearance o f Lesions. Lesions referred to electrode application were usually confined to segments and never widely disseminated over the whole crosssectional area of the sciatic nerve. Normal and pathologically altered regions could be distinguished easily in Groups 1 and 2, whereas in Group 3 differentiation was not as apparent in some cases. In regions classified as normal (Fig. 3), the mean nerve fiber diameter was 6.91 _+ 0.09 um and the mean nerve fiber density (neural cross-sectional area occupied by myelinated fibers) was 64.57% + 1.02%. Regions classified as pathologically altered appeared differently in the various rat groups. Ten days after electrode implantation (Group 1), the injured sectors exhibited fragmented myelin sheaths, myelin globules, signs of myelin phagocytosis, and a reduction of nerve fiber density due to an increase in connective tissue and edematous swelling. Within these sectors a small amount of relatively large normal-appearing nerve fibers could be observed (Fig. 4A). NormaI and pathologically altered sectors were distinguished on the basis of the mean nerve fiber density, which was reduced to 10.97% _+ 1.44% in damaged areas. Three weeks after electrode application (Group 2), damaged nerve segments usually contained several small nerve fibers with thin myelin sheaths and remnants of degenerated myelin sheaths (Fig. 4B)) 8 In this group, the mean nerve fiber density was 16.71% + 1.22% and the mean nerve fiber diameter was 4.84 _ 0.46 um in pathologically altered areas. Both criteria allowed a clear distinction (Fig. 5). Three months after the implantation procedure (Group 3), remnants of degenerated myelin sheaths in combination with small nerve fibers indicated pathologically altered areas in an advanced state of regenerJ. Neurosurg. / Volume 74/April, 1991
FIc. 4. Photomicrograph of semithin cross sections of rat sciatic nerves 8 mm distal to electrode application demonstrating typical segmentally localized pathological alterations. Calibration bar = 50 urn. A: Ten days after electrode implantation, a segment of the common peroneal branch exhibits distinct signs of nerve fiber degeneration. The surrounding nerve fibers are intact. B: At 3 weeks after electrode implantation, the circumscribed area (in the center) contains several small regenerating nerve fibers. C: At 3 months after electrode implantation, a segment of a small fascicle exhibits relatively large regenerating nerve fibers. The adjacent tibial branch does not show any pathological alterations.
ation in four rats (Fig. 4C). However, in two specimens of this group the altered segments still exhibited pronounced fibrosis and only a few regenerating nerve fibers were seen. 639
W. Girsch, et al. Histological Findings in Control Specimens Normal Sciatic Nerves. In sciatic nerves on the right side, which had not received any surgical manipulation, the mean diameter was 6.83 _+ 0.17 um and the mean percentage of fascicular area occupied by nerve fibers was 62.7% _ 1.71%. Sham-Operated Nerves. No signs of degeneration or regeneration could be detected in the six right sciatic nerves that had been exposed but did not receive electrodes. The mean diameter of myelinated fibers and the mean percentage of nerve fiber density in these specimens were 7.81 + 0.19 um and 70.11% + 1.93%, respectively. Crush-Injured Nerves. Contrary. to nerve segments altered by electrode implantation, no intact nerve fibers were seen in sciatic nerves of Group 1 crushed 10 days before. Three weeks after being crushed, the regenerating sciatic nerves exhibited nerve fibers with a mean diameter of 3.37 _+ 0.13 um and a mean nerve fiber density percentage of 20.31% _+ 3.31%. The corresponding data 3 months after crushing was 5.44 _+ 0.19 um and 49.82% _+ 1.32%, respectively. In Groups 2 and 3 the appearance of alterations resulting from electrode application generally corresponded well with that of the specimens of the crushed right sciatic nerves in Group 1 (Fig. 5). Discussion The results of the present study reveal that ringshaped stainless steel electrodes sutured to the epineurium of a peripheral nerve do not alter the morpholog-
ical aspect of the nerve to a great extent, provided that microsurgical techniques are carefully used. The nerve lesions observed in this investigation were clearly visible, but in general were circumscribed, not extensive, and clearly decreased over time.
Previous Studies In all previous studies dealing with the influence of cuff, coiled wire, or intraneurally implanted electrodes on nerve integrity, nerve fiber damage and myelin degeneration, or at least an increase in connective tissue due to electrode application was observed. ~-3'8'9"j1-16.23 Thus, at least minimal morphological alterations of nervous tissue seem to be an inevitable feature of functional electrical stimulation in peripheral nerves, even if lesions do not always occur. In some of the previous studies 2'3'~'~3-~5 an electrical current was applied to the nerves prior to histological examination. The question remained whether the influence of the electrical current or mechanical factors were responsible for the alterations observed in peripheral nerves after functional electrical stimulation. According to relevant publications, alterations of nervous tissue are obviously dependent on the intensity and the frequency of electrical current. The present study was undertaken in order to quantify morphological lesions caused by epineurial electrode application. Since no electrical current has been applied, lesions can only be due to mechanical factors, either the operative procedure or chronic mechanical irritation by the implanted electrodes. Arrangement of Lesions It may be surprising on first sight that lesions associated with electrode application were usually arranged in sectors or segments and were never widely disseminated over the whole cross-sectional area of the nerve. According to relevant publications about nerve injuriess there is no intermingling of fibers in the distal portion of a nerve trunk. Our findings provide an excellent explanation as to why mechanical stress to a segment of the nerve results in segmentally localized lesions at a more distal level. In the some cases, lesions at the electrode level were not followed by detectable alterations distal to the site of the electrodes. Thus, there is every reason to believe that in these cases the mechanical stress of electrode implantation did not cause Wallerian degeneration but resulted in segmental demyelination.~9
FlG. 5. Frequency histogram showing diameters of 250 regenerating (open bars) and an equal number of normalappearing (filled bars) myelinated nerve fibers found in the left sciatic nerve at the electrode level of a Group 2 rat (3 weeks after electrode placement). 640
Time Course of Lesions Both the number and extent of damaged areas decreased distinctly from Group 1 to Group 3, possibly due to full nerve regeneration in areas that had been only slightly damaged by the implantation procedure. Since the Group 3 animals had undergone electrode implantation before the two other investigation groups, an increase in the surgeon's experience is not responsiJ. Neurosurg. / Volume 74/April, 1991
Nerve lesions caused by epineural electrodes ble for the decrease in the number and extent of lesions from Group 1 to Group 3. The persistence of fibrous tissue and reduction in nerve fiber density in two Group 3 specimens can be referred either to mechanical stress by the implanted electrodes or to heavy damage during the surgical procedure prohibiting proper regeneration in the 3-month period. ~9 Although the sciatic nerves exhibited distinctly visible lesions 10 days after electrode implantation, in only two cases did the extent of altered areas exceed 12% of the total area of the nerve. Three months after electrode implantation, none of the damaged nerve segments covered more than 5%. Combining all specimens distal to the electrodes in Group 3 shows that, on average, more than 99% of the total area of neural tissue was intact. In addition, nearly all pathologically altered nerve regions in Groups 2 and 3 exhibited signs of nerve fiber regeneration.
Cause of Lesions The number and extent of damaged areas in Group 1 and the distinct decrease of these two parameters in Groups 2 and 3 lead to the conclusion that the observed alterations occur at implantation. Electrodes left in place over time probably neither lead to new alterations nor prohibit nerve fiber regeneration. No signs of degeneration were seen in the control nerves that had been exposed without electrode application. The procedure of suturing electrodes to the epineufium of a peripheral nerve seems to be the most important factor responsible for morphological alterations. This mechanical stress, probably due to manipulations with microsurgical instruments, is obviously not comparable to the crush injury of nerves on the control side. The crush lesions in the control nerves were, especially in Group 1, more severe than those observed in the reference group with electrode application.
Conclusions
Although no functional measurements have been performed in the present study, the extent of lesions and their decrease with time make long-term impairment of peripheral nerve function by the application of epineurially sutured electrodes extremely improbable. Of course, rat nerves may regenerate faster than human nerves, but the general ability to regenerate is the same in rats as in humans. Thus, the results obtained from rat nerves in this study should be transferable to human nerves in which electrodes are applied for "carousel stimulation." At the time of autopsy, nine of 140 electrodes were found to have become dislocated. All the other electrodes were fixed by soft connective tissue in their original position in contact with the epineurium. They would easily be removed by use of microsurgical techniques. Thus, epineurial electrode application seems to
J. Neurosurg. / Volume 74/April, 1991
be a method of fixing electrodes in a stable position to the nerve as close as possible without causing extensive and permanent alterations in nervous tissue. References
1. Agnew WF, McCreery DB, Yuen TGH, et al: Histologic and physiologic evaluation of electrically stimulated peripheral nerve: considerations for the selection of parameters. Ann Biomed Eng 17:39-60, 1989 2. Bowman BR, Erickson RC II: Acute and chronic implantation of coiled wire intraneural electrodes during cyclical electrical stimulation. Ann Biomed Eng 13:75-93, 1985 3. Breederveld RS: Electrical stimulation of motor nerves: results of animal experiment. Life Snpp Syst 2:183-188, 1984 4. Fodstad H: The Swedish experience in phrenic nerve stimulation. PACE 10:246-251, 1987 5. Glenn WWL, Holcomb WG, Hogan J, et al: Diaphragm pacing by radiofrequency transmission in the treatment of chronic ventilatory insufficiency. J Thor Cardiovasc Surg 66:505-520, 1973 6. Glenn WWL, Holcomb WG, Hogan JF, et al: Long-term stimulation of the phrenic nerve for diaphragm pacing, in Hambrecht FT, Reswick JB (eds): Functional Electrical Stimulation. New York: Marcel Decker, 1977, pp 97-112 7. Glenn WWL, Holcomb WG, McLaughlin AJ, et al: Total ventilatory support in a quadriplegic patient with radiofrequency electrophrenic respiration. N Engl J Med 286: 513-516, 1972 8. Happak W, Koller R, Girsch W, et al: Histological examination after epineural electrode application, in Wallinga W, Boom HBK, de Vries J (eds): Eleetrophysiologieal Kinesiology. Amsterdam: Excerpta Medica, 1988, pp 75-78 9. Herschberg PI, Sohn D, Agrawal GP, et al: Histologic changes in continuous, long-term electrical stimulation of a peripheral nerve. IEEE Trans Biomed Eng 14: 109-114, 1967 10. Holle J, Moritz E, Thoma H: Die Karussellstimulation, eine neue Methode zur Elektrophrenischen Langzeitbeatmung. Wien Kiln Wochenschr 86:23-27, 1974 11. Hughes GB, Bottomy MB, Dickins JRE, et al: A comparative study of neuropathologic changes following pulsed and direct current stimulation of mouse sciatic nerve. Am J Otolaryngol 1:378-385, 1980 12. Hughes GB, Bottomy MB, Jackson CG, et al: Myelin and axon degeneration following direct current peripheal nerve stimulation. A prospective controlled experimental study. Otolaryngol Head Neck Surg 89:767-775, 1981 13. Kim JH, Manuelidis EE, Glenn WWL, et al: Diaphragm pacing. Histopathologic changes in the phrenic nerves following long-term electrical stimulation. J Thor Cardiovasc Surg 72:602-608, 1976 14. Kim JH, Manuelidis EE, Glenn WWL, et al: Light and electron microscopic studies of phrenic nerves after longterm electrical stimulation. J Neurosurg 58:84-91, 1983 15. Rosenkranz D, Fenzl G, Holle J, et al: Influence of longterm low direct current on rat ischiatic nerves. Appl Neurophysio149:42-52, 1986 16. Rutten WLC, van Wier HJ, Put JHM, et al: Sensitivity, selectivity and bioacceptance of an intraneural multielectrode stimulation device in silicon technology, in Wallinga W, Boom HBK, de Vries J (eds): Electrophysiologteal Kinesiology. Amsterdam: Excerpta Medica, 1988, pp 135-139 641
W. Girsch, et al. 17. Schmalbruch H: Fiber composition of the rat sciatic nerve. Anal Rec 215:71-81, 1986 18. SchrOder JM: Altered ratio between axon diameter and myelin sheath thickness in regenerated neF,'e fibers, Brain Res 45:49-65, 1972 19. Sunderland S: Nerves and Nerve Injuries, ed 2. New York: Churchill Livingstone, 1978, pp 31-132 20. Thoma H, Gerner H, Girsch W, et al: lmplantable neurostimulators: the phrenie pacemaker-technology and rehabilitation strategies, in Wallinga W, Boom HBK. de Vries J (eds): Electrophysiologieal Kinesiology. Amsterdam: Exeerpta Mediea, 1988, pp 143-152 21. Waters RL, McNeal D, Perry' J: Experimental correction of footdrop by electrical stimulation of the peroneal nerve. J Bone Joint Surg (Am) 57:1047-1054, 1975
642
22, Young RF: Diaphragm pacing asan adjunct in respiratory insufficiency. Neurosurgery 2:43-46, 1978 23. Yuen TGH, Agnew WE, Bullara LA: Histopathological evaluation of dog sacral nerve after chronic electrical stimulation for micturition. Neurosurgery 14:449-455, 1984 Manuscript received April 1[, 1990. Accepted in final form September 21, 1990. This work was supported by the "Lorenz-Bohler Foundation" and the Austrian Research Foundations, FWF and FFF, the Ministu of Research and Science, and the Austrian National Bank. Address reprinl requests to." Werner Girsch, M.D., 2rid Surgical University Clinic, Spitalgasse 23, A-1090 Vienna, Austria.
J. Neurosurg. / Volume 74/April, 1991
