- Email: [email protected]
Diaphragm pacing
Diaphragm pacing Histopathological changes in the phrenic nerve following long-term electrical stimulation Phrenic nerves were obtained at autopsy from 7 patients with chronic ventilatory insufficiency. One or both nerves had been subjected to electrical stimulation to pace the diaphragm for about 2 months to 4 years. Sections of the nerves made from above, at the level of and below the site of application of the cuff electrode were studied microscopically. In most specimens, focal areas of demyelination were seen ranging in severity from swelling and fragmentation of isolated fibers to focal myelin destruction with phagocytic activity and total removal of myelin with fibrosis. Some axone loss was seen. The fact that the nerves from 2 patients, each stimulated for about 2 years, showed no changes suggests that injury to the nerve was not caused by the electrical stimuli but rather was secondary to the technique of application and fixation of the cuff electrode to the nerve.
Jung H. Kim, M.D., Elias E. Manuelidis, M.D., William W. L. Glenn, M.D., and Toshuhiro Kaneyuki, M.D., New Haven, Conn.
v 3 ince the successful application of diaphragm pacing to the treatment of ventilatory insufficiency of various origins, the effect of long-term electrical stimulation on the phrenic nerve has been of major concern to investigators. 1 - 3 Our lack of knowledge of the full consequences has been one of the major reasons for limiting the use of diaphragm pacing. Damage to the nerve, either electrical or mechanical, could result in permanent paralysis of the diaphragm which would be devastating in patients already in need of ventilatory assistance. Other than a brief account published elsewhere of two cases from our institution, there have been no reports on the histopathological changes in the phrenic nerve of patients treated with long-term diaphragm pacing. 1 ' 4 One report has been published on the results of long-term stimulation of the phrenic nerve in dogs. 2 This paper presents findings from histologic examination of the phrenic nerves of seven patients treated From the Departments of Pathology and Surgery, Yale University School of Medicine, New Haven, Conn. Supported by U.S. Public Health Service grants HL04651, HL14179, RR00125, and NS 05292-16 and by the Culpeper Foundation. Received for publication April 26, 1976. Accepted for publication June 30, 1976. 602
with diaphragm pacing over a period of several months to about four years. Patients and methods Subjects. All 7 patients required ventilatory support because of moderate-to-severe hypoventilation from various causes (Table I). Five suffered from chronic ventilatory insufficiency secondary to impairment of ventilatory control (Ondine's curse); these patients underwent diaphragm pacing for 8 to 12 hours nightly. The other 2 patients had a traumatic transection of the cervical cord and required total ventilatory support by continuous pacing; this was accomplished by stimulation of both phrenic nerves alternately for 12 hours each. Five patients showed marked clinical improvement during pacing, and the apparatus was apparently functioning well at the time of the patient's death. Electronic apparatus. The details of the electronic apparatus have been described in previous reports. 1 , 4 ' 5 Briefly, the apparatus consisted of a battery-operated radiofrequency transmitter with attached antenna and a passive receiver circuit connected to a silicone rubber-encased bipolar platinum electrode (electrode cuff) (Fig. 1). In Case 1, an amplitude-modulated radiofrequency transmitter tuned to 2.0 mHz was used and, in the other cases, a pulse-width modulated model.
Diaphragm pacing
603
Fig. 1A. Bipolar phrenic nerve electrode used in the cases reported herein. Electrical stimulation is through the cathode electrode placed cephalad on the nerve. The two platinum electrodes are fixed at right angles in the silicone rubber cuff. The internal diameter of the cuff is 5 mm. and the length of the cuff is 12 mm. Operative technique. This, too, has been described previously. 1, 4- 6 With the subject under local infiltrative anesthesia, bipolar platinum ribbon electrodes encased in a silicone rubber cuff were placed around the phrenic nerve lying over the scalenus anticus muscle. An exception was Case 2, in which the electrode was placed around the phrenic nerve in the chest at about the level of the aortic arch. Bilateral implantation was needed in 2 patients with quadriplegia (Cases 4 and 6). Special precautions were taken to minimize mechanical trauma to the nerve and the nutrient vessels during its isolation and placement within the electrode cuff. The response of the diaphragm to pacing was observed at the operating table. To allow the acute effects of the surgical procedure upon the nerve to subside, scheduled daily pacing was not started until 10 to 14 days postoperatively. Pathologic examination. The pertinent autopsy findings together with the clinical diagnoses are seen in Table I. Seven cases were evaluated histologically.* With the exception of the nerves from Case 4 , t the *The phrenic nerves from Cases 6 and 7 were kindly supplied by Drs. R. K. Tiwari and J. L. Dunn. tKindly supplied, in paraffin block, along with microscopic slides, by Dr. James Reiss.
Fig. IB. Electrode cuff and nerve removed at autopsy from Case 3. The silicone rubber covering the anterior surface of the nerve and one of the electrodes has been cut away revealing one of the platinum electrodes and the jacket of fibrous tissue surrounding the nerve.
6 04
The Journal of Thoracic and Cardiovascular Surgery
Kim et al.
Table I. Clinical and postmortem data Case No. 1 2
Age
Sex
40 5'/2
Clinical diagnosis
Placement of of electrode
Adequacy of pacing
Duration of pacing
M
Idiopathic
Left phrenic nerve
Yes
26 mo.
F
Encephalitis
Left phrenic nerve
Yes
11 mo., 19 days
3
55
F
?Leigh's disease
Left phrenic nerve
Yes
50 mo.
4
36
M
Bilateral
Yes
2 mo., 22 days
5
66
M
Quadriplegia secondary to fracture of second cervical vertebra Congestive heart failure
Left phrenic nerve
Yes
1 mo., 19 days
6 7
31 4
M F
Quadriplegia Idiopathic
Bilateral Left phrenic nerve
Yes Yes
20 mo. 32 mo.
Autopsy Acute myocardial infarction Acute and chronic encephalitis, RV dilatation RV dilatation and hypertrophy, congestive heart failure, pulmonary arterial arteriosclerosis Partial atelectasis of right lower lobe, moderate coronary atherosclerosis Bronchopneumonia, diffuse paracinar emphysema Purulent meningitis Old residua of mild brain-stem encephalitis
Legend: RV, Right ventricle.
Fig. 2. Case 1. Transverse section of phrenic nerve at the level of electrode showing thick capsular connective tissue (c) and microscopically unremarkable large (If) and small fascicles (sf). Note the loose epineurial fibroadipose tissue with several patent nutrient vessels (v). (Masson's trichrome stain, x30.)
stimulated phrenic nerves were removed with the electrode cuffs intact. The latter were carefully dissected away and the specimens immediately placed in 10 per cent neutral formalin. After fixation the nerves were sectioned into three segments, a proximal section, a middle section at the level of the cuff, and a distal section. Each segment was embedded in paraffin and cut at 10/A. The following stains were utilized: hematoxylin and eosin, Luxol fast blue,7 Bodian's silver,8 and Masson9 trichrome. Grossly, firm gray-white tissue was seen lining the electrode cuff. No gross changes were observed in the proximal and distal segments of the phrenic nerves. Microscopically, the most conspicuous changes were seen at the region of the cuff. The lining was composed of dense fibrous tissue of varying thicknesses (Fig. 2). In this fibrous capsule and in the underlying loose fibroadipose epineurial tissue there were occasional foci of round-cell infiltrates. In the epineurium, scattered hemosiderin-laden macrophages were present (Cases 2 and 6) as well as a few multinucleated foreign body giant cells containing birefringent material (Cases 2, 5, and 6). In contrast with these cases, the alterations in Case 7 were exceptionally severe. The fibrous capsule and the epineurium showed extensive regions of granulation tissue with a scattering of polymorphonuclear leukocytes and many mononuclear cells. In addition, there were also hemosiderin-laden mac-
Volume 72 Number 4 October, 1976
Diaphragm pacing
Fig. 3. Case 5. Transverse section of the phrenic nerve, at the level of the electrode, showing focal loss of myelin with flbrosis (arrowhead) and focal perineural thickenings (p) in large fascicle (If). In three small demyelinated fascicles (sf) fibrous tissue is seen. There is considerable fibrosis in the epineurium (e), and moderate round cell infiltrates are seen. (Luxol fast blue stain, x63.)
Fig. 4. A, Case 2. Longitudinal section of the phrenic nerve, at the level of the electrode, showing demyelination, globular swelling, and irregular fragmentation of preserved myelin. (Luxol fast blue stain, X256.) B, Case 4. Longitudinal sections of the phrenic nerve at the level of the electrode. Note many macophages (arrowheads) and patchy demyelination. Some nerve fibers show preserved myelin (in). (Masson's trichrome stain, x256.)C, Case 5. Longitudinal section of the phrenic nerve, at the level of the electrode, showing patchy demyelination and fibrosis (/) as well as scattered small foci of demyelination. (Luxol fast blue stain, x256.)
605
The Journal of Thoracic and Cardiovascular Surgery
6 0 6 Kim et al.
Fig. 5. A, Case 3. Longitudinal section of the phrenic nerve, at the level of the electrode, showing segmental demyelination (between two arrowheads) in a nervefiber.(Luxol fast blue stain, x256.) B, Case 6. Longitudinal section of the phrenic nerve, at the level of the electrode, showing normal-appearing myelinated nerve fibers. (Luxol fast blue stain, x256.) C, Case 7. Transverse section of the phrenic nerve, at the level of the electrode showing marked loss axons. Some of the preserved axons are marked with arrowheads. (Bodian's silver stain, x266.) rophages and foreign body giant cells. No changes were seen in the epineurium proximal and distal to the cuff. In Cases 5 and 7 a few fasciculi revealed focal thickening of the perineurium (Fig. 3), but in the other cases no such alterations were seen. There was no uniformity in the size and number of the phrenic nerve fascicles owing to the diversity of their branching and fusion. In general, the histologic changes in the fascicles occurred predominantly at the surface of the nerve and were most striking at the level of the electrode; the distal and proximal segments each showed, respectively, less severe changes. No such changes were noted at any level in Cases 1 and 6. In the other cases (Cases 2 , 3 , 4 , 5 , and 7), varying degrees of demyelination were detected at the level of the cuff (Fig. 3). All stages of demyelination were observed, ranging from swelling and fragmentation of isolated fibers (Fig. 4, A) to focal myelin destruction with phagocytic activity (Fig. 4, B) and total removal of myelin with fibrosis (Figs. 3 and 4, C). Thus, occasionally, small fascicles showed total demyelination and fibrous scarring. The demyelination was predominantly patchy in most cases (Fig. 3, arrow, and 4, C); however, an occasional isolated nerve fiber demonstrated a segmental type of demyelination (Fig. 5, A). In all cases examined there were many nerve fibers with well-preserved myelin (Fig. 5, B). Several foci of demyelination were seen in Cases 4 and 5, fewer in Case 2, and fewer yet in Case 3. In Case 7, in which extensive granulation tissue was found in the
epineurium, numerous foci of demyelination were seen as well as mononuclear infiltrates within the fascicles. In this case, even the preserved myelinated fibers demonstrated swelling and pallor. Some axonal loss was seen, although not all demyelinated fibers were so affected (Fig. 5, C). Discussion It has been demonstrated clinically that after an initial rise subsequent to application of the electrode, there is no progressive rise in the threshold to electrical stimulation and in some cases there is a decrease.4, 6 Most probably the initial rise is due to a buildup (fibrous capsule) of connective tissue within the electrode cuff. In canine experiments, the nerve action potential did not change during diaphragm fatigue induced by electrical stimulation of the phrenic nerve, and with rest full recovery from the fatigue was possible.10 The stability of the threshold on long-term pacing and the rapid recovery from fatigue suggest that the small electric stimulus, usually less than 2 Ma. applied intermittently, produces no permanent neural changes. This conclusion is supported by the absence of any microscopic lesion in the nerves from patients who were paced for nearly 2 years (Cases 1 and 6). Obviously another explanation must obtain for the changes observed in the 5 other cases. We believe there was direct injury to the nerve as the nerve cuff was applied. Histopathological changes in the nerves in our cases ranged from none to severe degeneration. In Case
Volume 72 Number 4 October, 1976
5, for example, the nerve was stimulated for no more than several 5 minute periods daily, yet it showed rather extensive degenerative changes. The demyelination was in all cases focal in character, and there were numerous myelinated fibers which did not reveal any histologic changes. In none of the cases could the severity of the degenerative changes be correlated with the duration of neural stimulation and with the actual current applied to the nerve. (One of the important features of the electronic circuit in our pacemaker design is its ability to measure precisely the current passing through the nerve.3) As a matter of fact, in the patient who required the highest current of any patient reported herein (5,120 Ma.) (Case 6, quadriplegia), no degenerative changes were observed in either of the two nerves. Data from experiments in our laboratory, performed with varying levels of current, suggest that there is less fatigue (therefore less possible injury to the nerve) when the lowest effective current is applied.11 Thus the histopathological findings in our cases seemed not to be due to electric stimuli, and it is likely that electrical stimuli such as have been applied in these patients produce no nerve damage. In animal experiments elsewhere, no microscopic abnormalities were found in either vagus12, 13 or splanchnic nerves14 subjected to long-term electrical stimulation. As regards phrenic nerves examined by still other investigators, degenerative changes of varying degrees were noted in stimulated as well as in nonstimulated (control) nerves.2 Our results bear out the theory that degenerative changes can be secondary to mechanical rather than to electrical factors. In placing the electrode cuff in these patients, manipulation of the phrenic nerve during its isolation from the surrounding tissue and placement within the cuff was unavoidable. Consequently, there was direct physical injury to the nerve. That some of the lesions in the fasciculi were located near the surface of the nerve favors this explanation. Ligation of major blood vessels supplying the sciatic nerve in rabbits produced only insignificant histopathological changes, probably because the longitudinal pathway supplying the nerve was preserved; however, slight mobilization (elevation) of this nerve at operation did result in extensive degenerative changes,15 presumably owing to direct injury to the nerve. Futher degenerative changes similar to those in our cases have been seen in peripheral nerves of patients with thromboangiitis obliterans,16 with diabetic mononeuropathy multiplex,17 and obliterative arteriosclerosis18 and in sciatic nerves of cats following compression by spring clips.19 In these cases the changes were explained on the basis of
Diaphragm pacing
607
ischemia. Vascular damage in some of our cases was evident by the presence of hemosiderin-laden macrophages. It appears that direct physical trauma to the nerve, at operation or thereafter because of malposition of the electrode cuff, and impairment of the blood supply to the nerve were the major causes of degenerative changes in our cases. Variation in the extent of such changes from nerve to nerve could be related to the individual neural blood supply pattern and/or the individual operative technique. It is of interest to note that even though there were focal degenerative changes in the phrenic nerve in 5 cases, there was no case without at least some diaphragm response shortly before death. This is in agreement with the finding of numerous unaltered myelinated fibers in all phrenic nerves examined. Case 5 presented exceptional features. The patient was subjected to only minimal phrenic nerve stimulation following placement of the electrode because of a marginal response to such stimulation at the time of operation. This poor response cannot be explained simply on the basis of the histologic findings. The degenerative changes noted in cross section, although somewhat more extensive than in most of the other cases, occurred in less than 25 per cent of the main fascicles. Since the most extensive ones were seen at the level of the electrode, changes must have developed following operation rather than before. One might speculate that the poor response of this patient at the time of operation was due to his severe metabolic acidosis, although other factors could have been involved. It has been shown in vitro that the action potential amplitude and conduction time in the phrenic nerve are altered in severe acidosis.20 Pathological findings in Case 7 were rather unusual. There was extensive granulation tissue in the epineurial and fibrous capsular tissue. It seems likely that these changes were due to an infection of undetermined origin, and because of inflammatory changes in and around the main fasciculus the role of physical trauma cannot properly be evaluated. Over the past 2 years experiments have been carried out in the laboratory which are aimed at the prevention of neural injury through the design of a less traumatic electrode. The results of these experiments will be reported in the near future. REFERENCES 1 Glenn, W. W. L., Holcomb, W. G., Gee, J. B. L., and Rath, R.: Central Hypoventilation: Long-Term Ventilatory Assistance by Radiofrequency Electrophrenic Respiration, Ann. Surg. 172: 755, 1970.
608
The Journal of Thoracic and Cardiovascular Surgery
Kim et al.
2 Hershberg, P. I., Sohn, D., Agrawal, G. P., and Kantrowitz, A.: Histologic Changes in Continuous, LongTerm Electrical Stimulation of a Peripheral Nerve, IEEE Trans. Biomed. Eng. Vol. BME-14: 109, 1967. 3 Sarnoff, S. J., Sarnoff, L. C , and Whittenberger, J. L.: Electrophrenic Respiration. VII. The Motor Point of the Phrenic Nerve in Relation to External Stimulation, Surg. Gynecol. Obstet. 93: 190, 1951. 4 Glenn, W. W. L., Holcomb, W. G., Hogan, J., Matano, I., Gee, J. B. L., Motoyama, E. K., Kim, C. S., Poirier, R. S., and Forbes, G.: Diaphragm Pacing by Radiofrequency Transmission in the Treatment of Chronic
11 12
13
CARDIOVASC.
14
5 Hogan, J. F., Holcomb, W. G., and Glenn, W. W. L.: Design of a Radiofrequency Pulse Duration Modulated System for Phrenic Nerve Pacing, Proc. Biomed. Symp. (San Diego), 14: 329, 1975. 6 Glenn, W. W. L., Holcomb, W. G., Shaw, R. K., Hogan, J. F., and Holschuh, K. R.: Long-Term Ventilatory Support by Diaphragm Pacing in Quadriplegia, Ann. Surg. 183: 560, 1976. 7 Kliiver, H., and Barrera, E.: A Method for the Combined Staining of Cells and Fibers in the Nervous System, J. Neuropathol. Exp. Neurol. 12: 400, 1953. 8 Luna, L. G.: Further Studies of Bodian's Technique, Am. J. Med. Technol. 30: 355, 1964. 9 Masson, P. J.: Trichrome Stainings and Their Preliminary Technique, J. Techn. Methods 12: 75, 1929. 10 Sato, G., Glenn, W. W. L., Holcomb, W. G., and Wuench, D.: Further Experience With Electrical Stimula-
15
Ventilatory Insufficiency,
SURG. 66: 505,
J. THORAC.
1973.
16 17 18 19 20
tion of the Phrenic Nerve: Electrically Induced Fatigue, Surgery 68: 817, 1970. Satoh, I., Fujii, Y., Kaneyuki, T., Hogan, J., Holcomb, W. G., and Glenn, W. W.L.: Totally Implantable Diaphragm Pacemaker, 1976. Submitted for publication. Bourde, J., Robinson, L. A., Suda, Y., and White, T. T.: Vagal Stimulation. I.: A Technique for Repeated Stimulation of the Vagus on Conscious Dogs, Ann. Surg. 171: 352, 1970. Palti, Y.: Stimulation of Internal Organs by Means of Externally Applied Electrodes, J. Appl. Physiol. 21: 1619, 1966. Fender, F. A.: Prolonged Splanchnic Stimulation, Proc. Soc. Exp. Biol. Med. 36: 396, 1937. Adams, W. E.: The Blood Supply of Nerves. II. The Effects of Exclusion of its Regional Sources of Supply on the Sciatic Nerve of the Rabbit, J. Anat. 77: 243, 1942. Barker, N. W.: Lesions of Peripheral Nerves in Thromboangiitis Obliterans, Arch. Intern. Med. 62: 271, 1938. Raff, M. C , Sangalang, V., and Asbury, A. K.: Ischemic Mononeuropathy Multiplex Associated With Diabetes Mellitus, Arch. Neurol. 18: 487, 1968. Eames, R. A., and Lange, L. S.: Clinical and Pathological Study of Ischaemic Neuropathy, J. Neurol. Neurosurg. Psychiatry. 32: 215, 1967. Denny-Brown, D., and Brenner, C : Lesion in Peripheral Nerve Resulting From Compression by Spring Clip, Arch. Neurol. Psychiatry. 52: 1, 1944. Ellis, F. R.: Some Effects of Pco2 and pH on Nerve Tissue, Br. J. Pharmacol. 35: 197, 1969.
Jung H. Kim, M.D., Elias E. Manuelidis, M.D., William W. L. Glenn, M.D., and Toshuhiro Kaneyuki, M.D., New Haven, Conn.
v 3 ince the successful application of diaphragm pacing to the treatment of ventilatory insufficiency of various origins, the effect of long-term electrical stimulation on the phrenic nerve has been of major concern to investigators. 1 - 3 Our lack of knowledge of the full consequences has been one of the major reasons for limiting the use of diaphragm pacing. Damage to the nerve, either electrical or mechanical, could result in permanent paralysis of the diaphragm which would be devastating in patients already in need of ventilatory assistance. Other than a brief account published elsewhere of two cases from our institution, there have been no reports on the histopathological changes in the phrenic nerve of patients treated with long-term diaphragm pacing. 1 ' 4 One report has been published on the results of long-term stimulation of the phrenic nerve in dogs. 2 This paper presents findings from histologic examination of the phrenic nerves of seven patients treated From the Departments of Pathology and Surgery, Yale University School of Medicine, New Haven, Conn. Supported by U.S. Public Health Service grants HL04651, HL14179, RR00125, and NS 05292-16 and by the Culpeper Foundation. Received for publication April 26, 1976. Accepted for publication June 30, 1976. 602
with diaphragm pacing over a period of several months to about four years. Patients and methods Subjects. All 7 patients required ventilatory support because of moderate-to-severe hypoventilation from various causes (Table I). Five suffered from chronic ventilatory insufficiency secondary to impairment of ventilatory control (Ondine's curse); these patients underwent diaphragm pacing for 8 to 12 hours nightly. The other 2 patients had a traumatic transection of the cervical cord and required total ventilatory support by continuous pacing; this was accomplished by stimulation of both phrenic nerves alternately for 12 hours each. Five patients showed marked clinical improvement during pacing, and the apparatus was apparently functioning well at the time of the patient's death. Electronic apparatus. The details of the electronic apparatus have been described in previous reports. 1 , 4 ' 5 Briefly, the apparatus consisted of a battery-operated radiofrequency transmitter with attached antenna and a passive receiver circuit connected to a silicone rubber-encased bipolar platinum electrode (electrode cuff) (Fig. 1). In Case 1, an amplitude-modulated radiofrequency transmitter tuned to 2.0 mHz was used and, in the other cases, a pulse-width modulated model.
Diaphragm pacing
603
Fig. 1A. Bipolar phrenic nerve electrode used in the cases reported herein. Electrical stimulation is through the cathode electrode placed cephalad on the nerve. The two platinum electrodes are fixed at right angles in the silicone rubber cuff. The internal diameter of the cuff is 5 mm. and the length of the cuff is 12 mm. Operative technique. This, too, has been described previously. 1, 4- 6 With the subject under local infiltrative anesthesia, bipolar platinum ribbon electrodes encased in a silicone rubber cuff were placed around the phrenic nerve lying over the scalenus anticus muscle. An exception was Case 2, in which the electrode was placed around the phrenic nerve in the chest at about the level of the aortic arch. Bilateral implantation was needed in 2 patients with quadriplegia (Cases 4 and 6). Special precautions were taken to minimize mechanical trauma to the nerve and the nutrient vessels during its isolation and placement within the electrode cuff. The response of the diaphragm to pacing was observed at the operating table. To allow the acute effects of the surgical procedure upon the nerve to subside, scheduled daily pacing was not started until 10 to 14 days postoperatively. Pathologic examination. The pertinent autopsy findings together with the clinical diagnoses are seen in Table I. Seven cases were evaluated histologically.* With the exception of the nerves from Case 4 , t the *The phrenic nerves from Cases 6 and 7 were kindly supplied by Drs. R. K. Tiwari and J. L. Dunn. tKindly supplied, in paraffin block, along with microscopic slides, by Dr. James Reiss.
Fig. IB. Electrode cuff and nerve removed at autopsy from Case 3. The silicone rubber covering the anterior surface of the nerve and one of the electrodes has been cut away revealing one of the platinum electrodes and the jacket of fibrous tissue surrounding the nerve.
6 04
The Journal of Thoracic and Cardiovascular Surgery
Kim et al.
Table I. Clinical and postmortem data Case No. 1 2
Age
Sex
40 5'/2
Clinical diagnosis
Placement of of electrode
Adequacy of pacing
Duration of pacing
M
Idiopathic
Left phrenic nerve
Yes
26 mo.
F
Encephalitis
Left phrenic nerve
Yes
11 mo., 19 days
3
55
F
?Leigh's disease
Left phrenic nerve
Yes
50 mo.
4
36
M
Bilateral
Yes
2 mo., 22 days
5
66
M
Quadriplegia secondary to fracture of second cervical vertebra Congestive heart failure
Left phrenic nerve
Yes
1 mo., 19 days
6 7
31 4
M F
Quadriplegia Idiopathic
Bilateral Left phrenic nerve
Yes Yes
20 mo. 32 mo.
Autopsy Acute myocardial infarction Acute and chronic encephalitis, RV dilatation RV dilatation and hypertrophy, congestive heart failure, pulmonary arterial arteriosclerosis Partial atelectasis of right lower lobe, moderate coronary atherosclerosis Bronchopneumonia, diffuse paracinar emphysema Purulent meningitis Old residua of mild brain-stem encephalitis
Legend: RV, Right ventricle.
Fig. 2. Case 1. Transverse section of phrenic nerve at the level of electrode showing thick capsular connective tissue (c) and microscopically unremarkable large (If) and small fascicles (sf). Note the loose epineurial fibroadipose tissue with several patent nutrient vessels (v). (Masson's trichrome stain, x30.)
stimulated phrenic nerves were removed with the electrode cuffs intact. The latter were carefully dissected away and the specimens immediately placed in 10 per cent neutral formalin. After fixation the nerves were sectioned into three segments, a proximal section, a middle section at the level of the cuff, and a distal section. Each segment was embedded in paraffin and cut at 10/A. The following stains were utilized: hematoxylin and eosin, Luxol fast blue,7 Bodian's silver,8 and Masson9 trichrome. Grossly, firm gray-white tissue was seen lining the electrode cuff. No gross changes were observed in the proximal and distal segments of the phrenic nerves. Microscopically, the most conspicuous changes were seen at the region of the cuff. The lining was composed of dense fibrous tissue of varying thicknesses (Fig. 2). In this fibrous capsule and in the underlying loose fibroadipose epineurial tissue there were occasional foci of round-cell infiltrates. In the epineurium, scattered hemosiderin-laden macrophages were present (Cases 2 and 6) as well as a few multinucleated foreign body giant cells containing birefringent material (Cases 2, 5, and 6). In contrast with these cases, the alterations in Case 7 were exceptionally severe. The fibrous capsule and the epineurium showed extensive regions of granulation tissue with a scattering of polymorphonuclear leukocytes and many mononuclear cells. In addition, there were also hemosiderin-laden mac-
Volume 72 Number 4 October, 1976
Diaphragm pacing
Fig. 3. Case 5. Transverse section of the phrenic nerve, at the level of the electrode, showing focal loss of myelin with flbrosis (arrowhead) and focal perineural thickenings (p) in large fascicle (If). In three small demyelinated fascicles (sf) fibrous tissue is seen. There is considerable fibrosis in the epineurium (e), and moderate round cell infiltrates are seen. (Luxol fast blue stain, x63.)
Fig. 4. A, Case 2. Longitudinal section of the phrenic nerve, at the level of the electrode, showing demyelination, globular swelling, and irregular fragmentation of preserved myelin. (Luxol fast blue stain, X256.) B, Case 4. Longitudinal sections of the phrenic nerve at the level of the electrode. Note many macophages (arrowheads) and patchy demyelination. Some nerve fibers show preserved myelin (in). (Masson's trichrome stain, x256.)C, Case 5. Longitudinal section of the phrenic nerve, at the level of the electrode, showing patchy demyelination and fibrosis (/) as well as scattered small foci of demyelination. (Luxol fast blue stain, x256.)
605
The Journal of Thoracic and Cardiovascular Surgery
6 0 6 Kim et al.
Fig. 5. A, Case 3. Longitudinal section of the phrenic nerve, at the level of the electrode, showing segmental demyelination (between two arrowheads) in a nervefiber.(Luxol fast blue stain, x256.) B, Case 6. Longitudinal section of the phrenic nerve, at the level of the electrode, showing normal-appearing myelinated nerve fibers. (Luxol fast blue stain, x256.) C, Case 7. Transverse section of the phrenic nerve, at the level of the electrode showing marked loss axons. Some of the preserved axons are marked with arrowheads. (Bodian's silver stain, x266.) rophages and foreign body giant cells. No changes were seen in the epineurium proximal and distal to the cuff. In Cases 5 and 7 a few fasciculi revealed focal thickening of the perineurium (Fig. 3), but in the other cases no such alterations were seen. There was no uniformity in the size and number of the phrenic nerve fascicles owing to the diversity of their branching and fusion. In general, the histologic changes in the fascicles occurred predominantly at the surface of the nerve and were most striking at the level of the electrode; the distal and proximal segments each showed, respectively, less severe changes. No such changes were noted at any level in Cases 1 and 6. In the other cases (Cases 2 , 3 , 4 , 5 , and 7), varying degrees of demyelination were detected at the level of the cuff (Fig. 3). All stages of demyelination were observed, ranging from swelling and fragmentation of isolated fibers (Fig. 4, A) to focal myelin destruction with phagocytic activity (Fig. 4, B) and total removal of myelin with fibrosis (Figs. 3 and 4, C). Thus, occasionally, small fascicles showed total demyelination and fibrous scarring. The demyelination was predominantly patchy in most cases (Fig. 3, arrow, and 4, C); however, an occasional isolated nerve fiber demonstrated a segmental type of demyelination (Fig. 5, A). In all cases examined there were many nerve fibers with well-preserved myelin (Fig. 5, B). Several foci of demyelination were seen in Cases 4 and 5, fewer in Case 2, and fewer yet in Case 3. In Case 7, in which extensive granulation tissue was found in the
epineurium, numerous foci of demyelination were seen as well as mononuclear infiltrates within the fascicles. In this case, even the preserved myelinated fibers demonstrated swelling and pallor. Some axonal loss was seen, although not all demyelinated fibers were so affected (Fig. 5, C). Discussion It has been demonstrated clinically that after an initial rise subsequent to application of the electrode, there is no progressive rise in the threshold to electrical stimulation and in some cases there is a decrease.4, 6 Most probably the initial rise is due to a buildup (fibrous capsule) of connective tissue within the electrode cuff. In canine experiments, the nerve action potential did not change during diaphragm fatigue induced by electrical stimulation of the phrenic nerve, and with rest full recovery from the fatigue was possible.10 The stability of the threshold on long-term pacing and the rapid recovery from fatigue suggest that the small electric stimulus, usually less than 2 Ma. applied intermittently, produces no permanent neural changes. This conclusion is supported by the absence of any microscopic lesion in the nerves from patients who were paced for nearly 2 years (Cases 1 and 6). Obviously another explanation must obtain for the changes observed in the 5 other cases. We believe there was direct injury to the nerve as the nerve cuff was applied. Histopathological changes in the nerves in our cases ranged from none to severe degeneration. In Case
Volume 72 Number 4 October, 1976
5, for example, the nerve was stimulated for no more than several 5 minute periods daily, yet it showed rather extensive degenerative changes. The demyelination was in all cases focal in character, and there were numerous myelinated fibers which did not reveal any histologic changes. In none of the cases could the severity of the degenerative changes be correlated with the duration of neural stimulation and with the actual current applied to the nerve. (One of the important features of the electronic circuit in our pacemaker design is its ability to measure precisely the current passing through the nerve.3) As a matter of fact, in the patient who required the highest current of any patient reported herein (5,120 Ma.) (Case 6, quadriplegia), no degenerative changes were observed in either of the two nerves. Data from experiments in our laboratory, performed with varying levels of current, suggest that there is less fatigue (therefore less possible injury to the nerve) when the lowest effective current is applied.11 Thus the histopathological findings in our cases seemed not to be due to electric stimuli, and it is likely that electrical stimuli such as have been applied in these patients produce no nerve damage. In animal experiments elsewhere, no microscopic abnormalities were found in either vagus12, 13 or splanchnic nerves14 subjected to long-term electrical stimulation. As regards phrenic nerves examined by still other investigators, degenerative changes of varying degrees were noted in stimulated as well as in nonstimulated (control) nerves.2 Our results bear out the theory that degenerative changes can be secondary to mechanical rather than to electrical factors. In placing the electrode cuff in these patients, manipulation of the phrenic nerve during its isolation from the surrounding tissue and placement within the cuff was unavoidable. Consequently, there was direct physical injury to the nerve. That some of the lesions in the fasciculi were located near the surface of the nerve favors this explanation. Ligation of major blood vessels supplying the sciatic nerve in rabbits produced only insignificant histopathological changes, probably because the longitudinal pathway supplying the nerve was preserved; however, slight mobilization (elevation) of this nerve at operation did result in extensive degenerative changes,15 presumably owing to direct injury to the nerve. Futher degenerative changes similar to those in our cases have been seen in peripheral nerves of patients with thromboangiitis obliterans,16 with diabetic mononeuropathy multiplex,17 and obliterative arteriosclerosis18 and in sciatic nerves of cats following compression by spring clips.19 In these cases the changes were explained on the basis of
Diaphragm pacing
607
ischemia. Vascular damage in some of our cases was evident by the presence of hemosiderin-laden macrophages. It appears that direct physical trauma to the nerve, at operation or thereafter because of malposition of the electrode cuff, and impairment of the blood supply to the nerve were the major causes of degenerative changes in our cases. Variation in the extent of such changes from nerve to nerve could be related to the individual neural blood supply pattern and/or the individual operative technique. It is of interest to note that even though there were focal degenerative changes in the phrenic nerve in 5 cases, there was no case without at least some diaphragm response shortly before death. This is in agreement with the finding of numerous unaltered myelinated fibers in all phrenic nerves examined. Case 5 presented exceptional features. The patient was subjected to only minimal phrenic nerve stimulation following placement of the electrode because of a marginal response to such stimulation at the time of operation. This poor response cannot be explained simply on the basis of the histologic findings. The degenerative changes noted in cross section, although somewhat more extensive than in most of the other cases, occurred in less than 25 per cent of the main fascicles. Since the most extensive ones were seen at the level of the electrode, changes must have developed following operation rather than before. One might speculate that the poor response of this patient at the time of operation was due to his severe metabolic acidosis, although other factors could have been involved. It has been shown in vitro that the action potential amplitude and conduction time in the phrenic nerve are altered in severe acidosis.20 Pathological findings in Case 7 were rather unusual. There was extensive granulation tissue in the epineurial and fibrous capsular tissue. It seems likely that these changes were due to an infection of undetermined origin, and because of inflammatory changes in and around the main fasciculus the role of physical trauma cannot properly be evaluated. Over the past 2 years experiments have been carried out in the laboratory which are aimed at the prevention of neural injury through the design of a less traumatic electrode. The results of these experiments will be reported in the near future. REFERENCES 1 Glenn, W. W. L., Holcomb, W. G., Gee, J. B. L., and Rath, R.: Central Hypoventilation: Long-Term Ventilatory Assistance by Radiofrequency Electrophrenic Respiration, Ann. Surg. 172: 755, 1970.
608
The Journal of Thoracic and Cardiovascular Surgery
Kim et al.
2 Hershberg, P. I., Sohn, D., Agrawal, G. P., and Kantrowitz, A.: Histologic Changes in Continuous, LongTerm Electrical Stimulation of a Peripheral Nerve, IEEE Trans. Biomed. Eng. Vol. BME-14: 109, 1967. 3 Sarnoff, S. J., Sarnoff, L. C , and Whittenberger, J. L.: Electrophrenic Respiration. VII. The Motor Point of the Phrenic Nerve in Relation to External Stimulation, Surg. Gynecol. Obstet. 93: 190, 1951. 4 Glenn, W. W. L., Holcomb, W. G., Hogan, J., Matano, I., Gee, J. B. L., Motoyama, E. K., Kim, C. S., Poirier, R. S., and Forbes, G.: Diaphragm Pacing by Radiofrequency Transmission in the Treatment of Chronic
11 12
13
CARDIOVASC.
14
5 Hogan, J. F., Holcomb, W. G., and Glenn, W. W. L.: Design of a Radiofrequency Pulse Duration Modulated System for Phrenic Nerve Pacing, Proc. Biomed. Symp. (San Diego), 14: 329, 1975. 6 Glenn, W. W. L., Holcomb, W. G., Shaw, R. K., Hogan, J. F., and Holschuh, K. R.: Long-Term Ventilatory Support by Diaphragm Pacing in Quadriplegia, Ann. Surg. 183: 560, 1976. 7 Kliiver, H., and Barrera, E.: A Method for the Combined Staining of Cells and Fibers in the Nervous System, J. Neuropathol. Exp. Neurol. 12: 400, 1953. 8 Luna, L. G.: Further Studies of Bodian's Technique, Am. J. Med. Technol. 30: 355, 1964. 9 Masson, P. J.: Trichrome Stainings and Their Preliminary Technique, J. Techn. Methods 12: 75, 1929. 10 Sato, G., Glenn, W. W. L., Holcomb, W. G., and Wuench, D.: Further Experience With Electrical Stimula-
15
Ventilatory Insufficiency,
SURG. 66: 505,
J. THORAC.
1973.
16 17 18 19 20
tion of the Phrenic Nerve: Electrically Induced Fatigue, Surgery 68: 817, 1970. Satoh, I., Fujii, Y., Kaneyuki, T., Hogan, J., Holcomb, W. G., and Glenn, W. W.L.: Totally Implantable Diaphragm Pacemaker, 1976. Submitted for publication. Bourde, J., Robinson, L. A., Suda, Y., and White, T. T.: Vagal Stimulation. I.: A Technique for Repeated Stimulation of the Vagus on Conscious Dogs, Ann. Surg. 171: 352, 1970. Palti, Y.: Stimulation of Internal Organs by Means of Externally Applied Electrodes, J. Appl. Physiol. 21: 1619, 1966. Fender, F. A.: Prolonged Splanchnic Stimulation, Proc. Soc. Exp. Biol. Med. 36: 396, 1937. Adams, W. E.: The Blood Supply of Nerves. II. The Effects of Exclusion of its Regional Sources of Supply on the Sciatic Nerve of the Rabbit, J. Anat. 77: 243, 1942. Barker, N. W.: Lesions of Peripheral Nerves in Thromboangiitis Obliterans, Arch. Intern. Med. 62: 271, 1938. Raff, M. C , Sangalang, V., and Asbury, A. K.: Ischemic Mononeuropathy Multiplex Associated With Diabetes Mellitus, Arch. Neurol. 18: 487, 1968. Eames, R. A., and Lange, L. S.: Clinical and Pathological Study of Ischaemic Neuropathy, J. Neurol. Neurosurg. Psychiatry. 32: 215, 1967. Denny-Brown, D., and Brenner, C : Lesion in Peripheral Nerve Resulting From Compression by Spring Clip, Arch. Neurol. Psychiatry. 52: 1, 1944. Ellis, F. R.: Some Effects of Pco2 and pH on Nerve Tissue, Br. J. Pharmacol. 35: 197, 1969.