Light and electron microscopic studies of phrenic nerves after long-term electrical stimulation

Transcription

1 J Neurosurg 58:84-91, 1983 Light and electron microscopic studies of phrenic nerves after long-term electrical stimulation JUNG H. KIM, M.D., ELIAS E. MANUELIDIS, M.D., WILLIAM W. L. GLENN, M.D., YOSHITAKA FUKUDA, M.D., DOUGLAS S. COLE, M.D., AND JAMES F. HOGAN, B.S.E.E. Sections of Neuropathology and Cardiothoracic Surgery, Department of Surgery, Yale University School of Medicine, New Haven, Connecticut V' Light and electron microscopic evaluations were carried out on canine phrenic nerves subjected to longterm electrical stimulation. A total of 34 stimulated and 19 control nerves were studied by light microscopy, and 10 stimulated and five control nerves were evaluated by electron microscopy. Except in a few cases in which a higher current was used, the current used for stimulation was between 1 and 2 ma. The pulse width was 150 #sec. The typical charge per pulse was 0.22 ~C and charge density per pulse #C/sq cm of real area. The total number of days of electrical stimulation in individual phrenic nerves ranged from 4 to 374. No morphological changes in the phrenic nerve that could be attributed to the electrical stimulation were observed by Fight or electron microscopic study. There were, however, two phrenic nerves cuffed with bipolar electrodes which showed focal demyelination at the electrode level, but these changes were caused by factors other than the electrical stimulation. The results of the studies have direct clinical implications to long-term stimulation of phrenic nerves. KEY WORDS phrenic nerve 9 diaphragm pacing 9 electrical stimulation neuroprosthesis 9 diaphragm fatigue F OR several years, electrical neural stimulation has been used successfully to restore or enhance the normal functions of various organs or to modify adverse nonphysiological activity. 11,12,22 The benefit of diaphragm pacing accomplished through electrical stimulation of the phrenic nerve, in patients with congenital or acquired ventilatory failure, has been well documented. 1~176 Despite successful clinical trials, it has been questioned whether phrenic nerve electrical stimulation damages the nerve itself when the stimulation is applied on a long-term basis. ~7 In previous studies with light microscopy on human phrenic nerves that had been subjected to long-term electrical stimulation, we found no changes that could be related directly to the electrical stimulation, a5 In this communication, we report the results of morphological studies done on canine phrenic nerves subjected to long-term electrical stimulation. Additional studies were carried out on the same group of animals for the purpose of evaluating the electrical parameters for diaphragm pacing 24 and noting morphological changes in the diaphragm in response to electrical stimulation of the phrenic nerve2 Materials and Methods Thirty-five adult mongrel dogs of either sex were used for this study. A total of 53 phrenic nerves were examined by light microscopy and 15 by electron microscopy. In the light microscopy study, 34 nerves had received stimulation and 19 served as controls, 15 of these having had no electrode cuff and four having had a cuff but no electrical stimulation. Of the 15 nerves examined by electron microscopy, 10 had been stimulated and five served as controls. The number of days of electrical stimulation accumulated by any nerve ranged from 4 to 374 days. The period of electrode implantation ranged from 273 to 986 days. The electronic apparatus used in these experiments has been fully described elsewhere. 2~ Briefly, the apparatus consists of a subcutaneously implanted battery-powered stimulator and platinum nerve cuff electrode. The amplitude of the stimulus applied to the nerve can be programmed from 0 ma (off) to 5 ma in 16 discrete steps via an external programmer. The neural electrodes are of two types, bipolar (Fig. 1A) and monopolar (Fig. 1B). When a monopolar neural electrode is used, the indifferent electrode is a separate metal plate. 84 J. Neurosurg. / Volume 58 / January, 1983

2 Phrenic nerves after long-term stimulation In these experiments, cathodal stimulation was always used. The current was typically about 1.5 ma except in a few cases in which it was higher, up to 5 ma. The pulse duration (width) was set at 150 #sec. The typical charge delivered to the nerve, therefore, was btc per pulse, using the formula Q = I t, where Q is coulomb, I is current, and t is time in seconds. The geometric area of the platinum electrode, whether bi- or monopolar, is 0.17 sq cm and the real area 0.20 sq cm (real sq cm). Consequently, the typical charge density used in these experiments was /~C/sq cm of real area per pulse. The pulse interval varied from 28 (36 Hz) to 140 mseconds (7 Hz). Wave forms used were either unidirectional biphasic or bidirectional (Lilly type). The pulse time repetition rate (respiratory rate) was 20/min, although in a few experiments a respiratory rate of 10/min was tried. The inspiration duration was set at 1.3 seconds. Most animals were paced with alternating periods of 6 weeks of stimulation and 6 weeks of rest, 24 and some received stimulation continuously for 52 weeks. Operative techniques for electrode implantation have been reported earlier. 24 To sample phrenic nerves for histological examination, animals were anesthetized with intravenous Pentothal (thiopental sodium). While the cardiorespiratory function was active, thoracotomy was performed. Short segments of phrenic nerve were quickly obtained from the proximal, middle (with electrode cuff), and distal portions of the nerve and were immediately immersed in 10% formalin for light microscopy and in 2% glutaraldehyde in 0.2N cacodylate buffer, ph 7.2, for electron microscopy. For light microscopy, the samples were further processed for dehydration and paraffin embedding. Sections 6/~ in thickness were stained with hematox- FIG. 1. Schematic drawings of a bipolar electrode completely encircling a phrenic nerve trunk (A), and a monopolar electrode encircling half of the circumference of the nerve trunk (B). ylin and eosin (H & E), Masson's trichrome, 19 Bodian, as,32 and Luxol fast blue26 For electron microscopy, the fixed tissue was osmicated, dehydrated, and impregnated with Epon. Semi-thin sections (1 # thick) were stained with toluidine blue, and thin sections were made in selected blocks. They were stained with lead citrate and uranyl acetate. Control Nerves Results The phrenic nerves in our canine model were seen to be composed of a single nerve fascicle surrounded FIG. 2. Experiment H & E, x 39. A: Section of the left phrenic nerve at the electrode level after stimulation with a bipolar electrode for 126 days at 26 Hz. The epineurial fibroadipose layer is completely surrounded by a thick fibrous capsule (arrows), but the nerve fascicle itself is histologically unremarkable. B: Section of the right phrenic nerve at the electrode level after stimulation with a monopolar electrode for 154 days at 27 Hz. A band of fibrous tissue (arrow) has developed at the lower margin facing the electrode. The nerve fascicle is histologically unremarkable. J. Neurosurg. / Volume 58 / January,

3 J. H. Kim, et al. FIG. 3. A: Experiment Section of the left phrenic nerve at the electrode level after stimulation with a bipolar electrode for 201 days at 27 Hz. Compression of the left half of the nerve fascicle (Nf) is evident (arrow) by a nodular protrusion (arrowhead) of the fibrous capsule observed at the left lower corner. Masson's trichrome, 27. B: Experiment Section of the right phrenic nerve at the electrode level. Nerve stimulated with a bipolar electrode for a total of 186 days at 27 Hz. A focus of demyelination (large arrow) is seen at the right center of the field. Several macrophages (small arrows) are also present at the lower center between the perineurium (curved arrow) and the demyelinated focus. Luxol fast blue, x 770. by abundant epineurial fibroadipose tissue. The fibrous tissue in the epineurium was in general scanty, but a few nerves had irregular bands of thick fibrous tissue closely apposed to the perineurium. On transverse section, individual nerve fascicles were seen to consist of myelinated nerve fibers sprinkled with scattered groups of unmyelinated nerve fibers. In areas corresponding to myelin-negative regions, with Bodian silver staining there were barely visible silverpositive structures compatible with axons of small size. The distribution and amount of unmyelinated nerve fibers varied greatly from sample to sample. The control nerves with a non-stimulating electrode cuff showed essentially the same histology as did the nerves without a cuff, but had a layer of dense fibrous tissue in the epineurium facing the electrode. Stimulated Nerves Like the control nerves with unactivated electrode cuffs, stimulated nerves developed a layer of dense fibrous tissue in the epineurium underneath the electrode cuff. In the samples where the electrode was bipolar, completely encircling the nerve trunk, dense fibrous tissue encapsulation in the epineurium was always seen at the electrode level (Fig. 2A). In monopolar applications, only that area of epineurium FIG. 4. Experiment Section of the left phrenic nerve at the electrode level after stimulation with a monopolar electrode for 374 consecutive days at 26 Hz A: Bodian stain demonstrates no loss of axons. B: Masson's trichrome stain also fails to reveal any foci of fibrosis or demyelination. 86 J. Neurosurg. / Volume 58 / January, 1983

4 Phrenic nerves after long-term stimulation FIG. 5. Experiment Section of the left phrenic nerve at the electrode level after stimulation with a monopolar electrode for 368 consecutive days at 24 Hz. The transverse section of the nerve shows no areas of demyelination or architectural abnormalities. Luxol fast blue, x 158. facing the electrode cuff developed a layer of dense fibrous tissue (Fig. 2B). In the epineurium of most of the stimulated nerves, there was an abundance of fibroadipose tissue between the nerve fascicle and the dense fibrous capsular tissue. In some nerves, there was a questionable increase in the amount of fibrous tissue in the epineurial adipose layer over that in control nerves; because of the great variation in the amount of fibrous tissue present, even in the control group, a quantitative differentiation was not possible. Other occasional observations in the epineurium or in the fibrous capsular tissue of stimulated nerves were hemosiderin-laden macrophages and/or foreign-body giant cells containing suture material. Such changes, when seen, were confined to the electrode level. Rarely, granulation tissue involved the epineurium, and was frequently sprinkled with chronic inflammatory ceils. Blood vessels in the epineurium were patent. In two nerves with bipolar electrodes, focal loss of myelin was evident at the level of the nerve cuff (Fig. 3). Underneath the cuff in one, the nerve fascicle was partially compressed by nodular thickening of the fibrous capsule (Fig. 3A), and in the other nerve there was a build-up of markedly thickened fibrous tissue close to the perineurium. A few foamy macrophages were present near the demyelinated focus (Fig. 3B). There were also numerous hemosiderin granules in the epineurium of the same nerve, which attests to the presence of vascular trauma. There was, however, no obvious demyelination at the distal or proximal levels of both nerves. The perineurium in individual nerves was more or less uniform in thickness, and showed no significant focal or generalized thickening or other morphological alteration. With the exception of the nerves just described, the stimulated nerves had histologically unremarkable nerve fibers. More specifically, there was no evidence of patchy fibrosis or loss of myelin or axons on special staining (Figs. 4, 5, and 6B), of unusual proliferation of Schwann cells or macrophages, or of thickening of axons on Bodian staining. If one nerve in a given animal was stimulated, it was carefully compared with the contralateral unstimulated nerve. In no animal was there a significant histological difference between the two nerves (Fig. 6A and B). Electron Microscopy Both control and stimulated nerves showed similar ultrastructural morphology and are described to- FIG. 6. Experiment A: Section of the left phrenic nerve, when no electrode was applied to the nerve. B: Section of the right phrenic nerve at the electrode level after stimulation with a bipolar electrode for 172 days at 27 Hz. There is no obvious morphological difference between the two nerves. Scattered small bundles of unmyelinated nerves (darkfoci) are seen in both nerves. Masson's trichrome, 122. J. Neurosurg. / Volume 58 / January,

5 J. H. Kim, et al. FIG. 7. Experiment Section of the left phrenic nerve at the electrode level after stimulation with a monopolar electrode for 190 days at 27 Hz. A collagen pocket is shown (arrow). Unmyelinated axons and Schwann-cell cytoplasm are unremarkable. Electron microscopy, 26,770. gether. Moderate artifactual changes caused by the immersion technique used for tissue fixation were present in some control and stimulated nerves. The changes were compression artifacts, separation of Schmidt-Lanterman incisures, and condensation of the axoplasm, among others. In certain nerves, sec- tions from the middle (with electrode cuff) and distal levels only were studied. The perineurium was intact, composed of layers of closely apposed perineurial cells. No increase in the amount of collagen fibrils was appreciated among the perineurial cells in stimulated nerves, the endoneurium of which also failed to demonstrate an appreciable increase in collagen fibrils. There were abundant unmyelinated nerve fibers arranged in large groups. These fibers were free of degenerative changes, and Schwann cells also were unremarkable. Rare collagen pockets ensheathed by a Schwann-ceU process of unmyelinated nerve bundles were present in certain control and stimulated nerves (Fig. 7). The axons were well preserved, and no abnormality was appreciated in the number and distribution of subceuular organelles, including neurofilaments and mitochondria. No axons had accumulations of 10-nm neurofilaments, mitochondria, or degenerate structures (Fig. 8). The myelin sheath of myelinated nerve fibers was unremarkable, the periodicity of the myelin being well maintained. There were, however, rare intra-axonal protrusions of the adaxonal Schwann-cell sheath in otherwise normal fibers of both control and stimulated nerves (Fig. 8). No obvious difference in the frequency of the protrusions was seen between the two groups, although no quantitative estimation was attempted. Vessels in the endoneurium were unremarkable (Fig. 9). Rarely FIG. 8. A: Experiment Section of the right phrenic nerve at the electrode level after stimulation with a monopolar electrode for 139 days at 24 Hz. A small protrusion of adaxonal Schwann-cell sheath (arrow) is present in the axon at the right upper comer. All the other axons in this figure are unremarkable. No obvious increase in the amount of collagen fibrils is present among nerve fibers. Electron microscopy, B: Experiment Section of the right phrenic nerve at the electrode level after stimulation with a monopolar electrode for 154 days at 29 Hz. A serpentine protrusion of adaxonal Schwann-cell sheath (arrow) is present in the axon. The myelin sheath and axon are otherwise unremarkable. A small "myelin figure" (arrowhead) observed in the right lower field may represent an artifactual change. Electron microscopy, 16, J. Neurosurg. / Volume 58 / January, 1983

6 Phrenic nerves after long-term stimulation were mast cells and fibroblasts encountered in the endoneurium of both groups. Discussion From results of our current study, it is apparent that electrical stimuli, when properly applied, do not injure the phrenic nerve. This conclusion is in agreement with our findings from stimulation of human phrenic nerves. No changes on histological study were seen in nerves of two patients who had received diaphragm pacing for almost 2 years; and in nerves of other patients, there were no morphological changes that could be correlated with the duration of the application or with the mode of the electrical stimulation? ~ Most morphological changes seen in our series of human phrenic nerves were compatible with changes secondary to ischemia, and to mechanical trauma related to the application of a bipolar neural electrode to the phrenic nerve. Some changes, axonal loss in particular, were possibly the result of the lesion in the central nervous system which necessitated the diaphragm pacing. Other investigators have also reported a lack of morphological changes in human phrenic nerves subjected to long-term stimulation. 2~ Although a case was reported of an infant whose stimulated phrenic nerves showed considerable irregularity in axon diameter on Bodian staining, routine H & E and myelin stain study did not demonstrate any abnormalities. 17 Also noted in this same case was the disappearance of the axons in the nerve fibers and in fine nerve twigs of the diaphragm muscle, yet there was no typical neurogenic atrophy of the muscle. Authors of the report considered the axonal changes to be due to the electrical stimulation, and attempted to explain them on the basis of the "dying back" phenomenon. Throughout our evaluation of stimulated canine phrenic nerves, we have never encountered any abnormal axons. Bodian silver stain, known to be very capricious in achieving a good stain, 18,32 may not stain well axons of small caliber such as those in fibers of unmyelinated nerve and in fine nerve branches in muscle. In addition, the thickness of axons so stained varies according to the silver impregnation time, 21 for which reason it is not unusual to find considerable difference in the diameter of axons of the same nerve sectioned and stained on different occasions. Thus, assessment of the caliber of normal peripheral nerve axons in a given section is not reliables A further obstacle to accurate assessment of axonal caliber in man is the diversity in size (a range of 2.0 to 14.2/~) of human phrenic nerve fibers? 3 On electron microscopic study of the stimulated phrenic nerves, no axonal degeneration was apparent; that is, there was no accumulation of degenerate organelles or neurofilaments such as are typically seen in most "dying back" phenomena. A defect in the axoplasmic transport is postulated as a cause of the "dying back" phenomenon, but it has been well established that FIG. 9. Experiment Section of the right phrenic nerve after stimulation with a monopolar electrode for 154 days at 29 Hz. Endothelial cells in the endoneurial capillary vessel fail to demonstrate any degenerative changes. Arrows indicate junctions between endothelial cells. Electron microscopy, x repetitive electrical stimulation produces no effect on the axoplasmic transport mechanism. 14 Long-term electrical stimulation has also been shown by others not to damage other peripheral nerves, a,a,21,27 For our experiments, the typical charge per pulse was #C, and charge density 1.125/~C/ real sq cm/pulse, which was substantially lower than the neural damage thresholds for the cerebellum. 4,28,a4 The low level of electrical charge to the tissue may have been a major reason for the negative histological findings in our experiment. Another significant factor may be that, unlike the central nervous system, peripheral nerves are in general well ensheathed by the abundant epineurial fibroadipose tissue and dense perineurium, which conceivably protect to some extent against compression and other mechanical trauma, or against thermal or electrochemical toxic injury which electrical stimulation might produce. The perineurium is thought to be an efficient barrier against certain noxious substances outside the nerves? 6 It is interesting to note in this same group of animals that, although stimulation at a high frequency such as used in our experiments (33 Hz) rapidly brings on fatigue in the diaphragm muscle (and as a consequence a decrease in the tidal volume 2~) and produces severe morphological changes in the muscle fibers2 such stimulation does not injure the phrenic nerve. High-frequency stimulation causes a rapid decrease in the end plate potentials at the neuromuscular junction but no change in the duration or strength of nerve action potentials along the nerve fibers, even in cases where no diaphragm response was observed. 29 Thus, it appears there is a significant interference with the transmission of impulses across the neuromuscular J. Neurosurg. / Volume 58 / January,

7 J. H. Kim, et al. junction, especially when the current is of a high frequency. Of two nerves in which we observed focal demyelination at the level of the electrode cuff, one was compressed by a nodular thickening of the capsular fibrous build-up underneath the electrode. The other phrenic nerve also showed a markedly thickened fibrous capsule, in and around which there were many hemosiderin granules. These changes are similar to those in some human phrenic nerves studied earlier, 15 and do not appear related to the electrical stimulation. Both nerves had been cuffed with a bipolar electrode. Because an electrode of this type completely encircles the nerve trunk, some trauma to the nerve or its vessels during its application is not unlikely, and a further danger is chronic compression from capsular fibrous tissue developing underneath the electrode. It is of interest to find in these two cases no appreciable demyelination of the nerve either at the proximal or distal level. The absence of demyelination at other levels is similar to that observed in the natural median or ulnar nerve compression syndrome of guinea pigs. In mild cases of the syndrome, one can only observe local demyelination under the compressing band, with remyelination in the distal part of the nerves2 We are aware of no other report describing ultrastructural studies of electrically stimulated peripheral nerves. In our electron microscopic studies of selective nerves, no obvious axonal or myelin degeneration was observed and endoneurial vessels were intact. There is one report, however, which mentions severe vascular changes in the electrically stimulated cerebellar cortex. 1 Of particular interest in our study of the phrenic nerve is the presence in both control and stimulated nerves of rare intra-axonal protrusions of the adaxonal Schwann-cell sheath. The significance of the structure is obscure, but some regard it as the morphological counterpart of Schwann cells' phagocytic activity to sequestrate and remove any effete organelles from the axon? ~ It has been seen in various axonal degenerative disorders) ~ and the structure is also known to occur in normaf,6.3~ or aged :~1 animals. Collagen pockets observed in both stimulated and control nerves have also been reported in nerves with neuropathic disorders and in nerves from normal individuals. 23 Both the collagen pocket and intra-axonal protrusions of the Schwann-cell sheath are numerically insignificant, and no positive correlations could be made between the presence of the structures and the degree of electrical stimulation among the nerves studied. Whether the change is related to aging in particular animals cannot be determined because of the lack of information on age in our experimental animals. In conclusion, long-term electrical stimulation of the peripheral nerve of sufficient intensity to result in an appropriate functional response in the effector organ induces no appreciable neural damage as observed at light and electron microscopic levels. Acknowledgments The authors would like to thank Kathryn Jeffrey, Rachel Ardito, Elizabeth Mullaly, and Albert Coritz for their excellent technical assistance. References 1. Agnew WF, Yuen TGH, Pudenz RH, et al: Electrical stimulation of the brain. IV. Ultrastructural studies. Surg Neuroi 4: , Berthold CH: Ultrastructure of the node-paranode region of mature feline ventral lumbar spinal-root fibres. Acta Soc Med Upsal 73 (Suppi 9):37-70, Bourde J, Robinson LA, Suda Y, et al: Vagal stimulation: I. A technic for repeated stimulation of the vagus on conscious dogs. Ann Surg 171: , Brown WJ, Babb TL, Soper HV, et al: Tissue reactions to long-term electrical stimulation of the cerebellum in monkeys. J Neurosurg 47: , Ciesielski TE, Fukuda Y, Glenn WWL, et al: Response of the diaphragm muscle to electrical stimulation of the phrenic nerve. A histochemical and ultrastructural study. J Neurosurg 58:92-100, Collins GH, Webster HdeF, Victor M: The ultrastructure of myelin and axonal alterations in sciatic nerves of thiamine deficient and chronically starved rats. Aeta Neuropathol 3: , Duncan D: A relation between axone diameter and myelination determined by measurement of myelinated spinal root fibers. J Comp Neurol 60: , Fender FA: Prolonged splanchnic stimulation. Proc Soc Exp Biol Med 36: , Fullerton PM, Gilliatt RW: Median and ulnar neuropathy in the guinea-pig. J Neurol Neurosurg Psychiatry 30: , Glenn WWL: Diaphragm pacing: present status. PACE 1: , Glenn WWL, Hogan JF, Phelps ML: Ventilatory support of the quadriplegic patient with respiratory paralysis by diaphragm pacing. Surg Clin North Am 60(5): , Hambrecht FT: Applications of neural control in humans. Ann Biomed Eng 8: , Heinbecker P, Bishop GH, O'Leary JL: Functional and histologic studies of somatic and autonomic nerves of man. Arch Neuroi Psychiatry 35: , Ignelzi RJ, Nyquist JK: Observations on fast axoplasmic transport in peripheral nerve following repetitive electrical stimulation. Pain 7: , Kim JH, Manuelidis EE, Glenn WWL, et al: Diaphragm pacing. Histopathologic changes in the phrenic nerve following long-term electrical stimulation. J Thorac Cardiovasc Surg 72: , Kliiver H, Barrera E: A method for the combined staining of cells and fibers in the nervous system. J Neuropathol Exp Neurol 12: , Liu HM, Loew TM, Hunt CE: Congenital hypoventilation syndrome: a pathologic study of the neuromuscular system. Neurology 28: , Luna LG: Further studies of Bodian's technique; with special emphasis on the impregnating and reducing solutions. Am J Med Technol 30: , Masson P: Some histological methods; trichrome stainings and their preliminary technique. J Teehnol Meth 90 J. Neurosurg. / Volume 58 / January, 1983

8 Phrenic nerves after long-term stimulation 12:75-90, McMichan JC, Piepgras DG, Gracey DR, et al: Electrophrenic respiration. Report of six cases. Mayo Clin Proc 54: , l. McNeal DR, Waters R, Reswick J: Experience with implanted electrodes. Neurosurgery 1: , t Nashold BS Jr: Overview of neuroaugmentation. Neurosurgery 1: , Ochoa J: The unmyelinated nerve fibre, in Landon DN (ed): The Peripheral Nerve. London: Chapman and Hall, 1976, pp t Oda T, Glenn WWL, Fukuda Y, et al: Evaluation of electrical parameters for diaphragm pacing: an experimental study. J Surg Res 30: , Oda T, Hogan JF, Glenn WWL, et al: A totally implantable diaphragm pacemaker for experimental studies: effect of stimulating current level on diaphragm fatigue, in Friedman EA (ed): Proceedings of the Second Meeting of the ISAO. Cleveland: International Society for Artificial Organs, 1979, pp , Oldfors A, Johansson BR: Barriers and transport properties of the perineurium. An ultrastructural study with 125I-labeled albumin and horseradish peroxidase in normal and protein-deprived rats. Acta Neuropathol 47: , Palti Y: Stimulation of internal organs by means of externally applied electrodes. J Appi Physiol 21: , Pudenz RH, Agnew WF, Yuen TGH, et al: Electrical stimulation of brain. Light and electron microscopy studies, in Hambrecht FT, Reswick JB (eds): Functional Electrical Stimulation. New York: Marcell Dekker, 1977, pp Sato G, Glenn WWL, Holcomb WG, et al: Further experience with electrical stimulation of the phrenic nerve: electrically induced fatigue. Surgery 68: , Spencer PS, Schaumburg HH: Pathobiology of neurotoxic axonal degeneration, in Waxman SG (ed): Physiology and Pathobiology of Axons. New York: Raven Press, 1978, pp Thomas PK, King RHM, Sharma AK: Changes with age in the peripheralnerves of the rat. An ultrastructural study. Aeta Neuropatho152:1-6, Weinzimer SH: A modification of the Bodian stain for neurofibrils. Lab Med 5:23-24, Young RF: Diaphragm pacing as an adjunct in respiratory insufficiency. Neurosurgery 2:43-45, Yuen TGH, Agnew WF, Bullara LA, et al: Histological evaluation of neural damage from electrical stimulation: considerations for the selection of parameters for clinical application. Nenrosurgery 9: , 1981 Manuscript received January 4, Accepted in final form August 3, This work was supported by USPHS Grant 2 R01 HL 0465 t, the Charles E. Culpeper Foundation, and a gift from Mrs. Jane Fetter. Address reprint requests to: J. H. Kim, M.D,, Section of Neuropathology, Department of Surgery. Yale University School of Medicine, 333 Cedar Street. Box 3333, New Haven, Connecticut J. Neurosurg. / Volume 58 / January,