U.S. Army Medical Department, Office of Medical History
Skip Navigation, go to content







AMEDD MEDAL OF HONOR RECIPIENTS External Link, Opens in New Window






Chapter XII








Of the 174,296 battle injuries in the American Expeditionary Forces, 1 there were under treatment in military hospitals in the United States on May 1,1919, 2,347 patients with 2,707 nerve injuries. The total number of nerve injuries could not accurately be determined in view of the fact that in many instances patients suffering from such injuries were admitted to hospital for tither conditions with which the nerve injuries were associated. However, it was estimated at the time in question that 30 percent of the patients with nerve injuries either were on furlough or had been discharged. Taking this number into consideration, as well as the comparatively small number of patients with nerve injuries who inadvertently had been sent to hospitals other than those designated for the purpose, the total number of peripheral nerve injuries was estimated as being 3,500, or about 2 percent of the total number if battle injuries.


Relatively few operations for the repair of peripheral nerve injuries were performed in the American Expeditionary Forces; consequently, in anticipation of the arrival of such cases in the United States, certain general hospitals throughout the country were designated for their special care.2 In each of these hospitals, known as peripheral nerve centers, carefully selected, experienced neurosurgeons had been placed, and in each there were well organized physiotherapy and educational departments. In passing, it may be added that, although there were 12 of these peripheral nerve centers, the major portion of neurosurgical operations were performed in six of them. On January 29, 1919, the Surgeon General appointed a peripheral nerve commission. 3 whose function it was to study and correlate the cases of peripheral nerve injuries. To make possible such study and correlation there was issued to each center a prescribed form, the peripheral nerve register, upon which was reoorded the physical findings. Duplicates of these registers were deposited in the Surgeon General's Office upon the discharge of the patients concerned. 4

With an organization thus effected in the peripheral nerve center for the management of peripheral nerve lesions, under conditions approaching the ideal, it was hoped that this organization might be continued for the determination of end results under equally ideal conditions. This hope, however, was not realized. During 1919, the supervision of neurosurgical cases, remaining under observation and seemingly not requiring operative treatment, was

a The facts contained herein are based, in the main, on "A General Discussion of the Operative Treatment and the Results in Three Thousand Five Hundred Peripheral Nerve Injuries of the American Expeditionary Forces," by L.ieut.Charles H. Frazier, M. C., read before the International Surgical Society, London, July 19, 1923.


transferred to the United States Public Health Service, under the Bureau of War Risk Insurance.5 In 1921, a further change in supervision was made. The Veterans' Bureau now took over, from the Public Health Service, the care of such patients still under treatment, or on a commpensable status.6 Thus, to follow a certain number of the nerve injury patients, it was necessary to deal with them, at different times, through three Government agencies.


When comparing one series of statistics with another, or in correlating any individual series of operations from the standpoint of end results, the time element must be taken into consideration. It is generally conceded that the sooner the operation is performed after the injury the better the prognosis, and it has been stated that the results of nerve suture performed within the first 2 months after iniurv are better than those between 12 and 24 months or later. An attempt to confirm this statement by comparison of statistics in this series, however, failed. It was a general practice in the periphleral nerve centers to postpone operation until three months after the wound had healed, and as the majority of the nerve injury cases had infected wounds on their arrival at the hospitals, the time of operation was deferred from four to eight months. In a series of 400 cases in General Hospital No. 11, in 31.5 percent the operation was deferred to the end of the fourth month, in 41 percent to the end of the fifth or sixth month, and in 15.5 percent to the end of the eighth month (Table 9).

TABLE 9.- Data concerning time of operation in a series of 400 cases of peripheral nerve injury

To wait for spontaneous regeneration longer than the sixth month is not considered justifiable, although it is well known that the first clinical signs of spontaneous regeneration may be deferred much longer. One case of musculospiral paralysis was observed in which the first sign of the return of voluntary motion did not appear until 21 months after the injury.

While the majority of the operations in the peripheral nerve centers were performed between October, 1918, and August, 1919, that is within a year of the injury, a not inconsiderable number of nerve sutures were performed at later dates, mostly on patients who, in times of overcrowding, had not passed through the peripheral nerve centers. These patients with their untreated lesions were discovered as they appeared for reexamination and adjustment of compensation.


During the period of reconstruction one relied upon the various Signs of recovering function to determine whether spontaneous regeneration was in process or not, such as contraction of the field of sensory loss, the return of muscle tone, the changing electrical reactions, and the advancement of Tinel's sign. The latter, however, did not prove to be an infallible guide. Unfortunately, it was not until 1922 that a more accurate clinical method of determining the early recovery of a degenerated nerve was elaborated.7 With the aid of a specially constructed chronomyometer, the nerve-muscle complex was found to disappear as early as the fifth day after the muscle was cut; it begins to appear from one to six months before a faradic response, from one week to three months before voluntary contractions and from zero to five months before contraction of the anesthetic area. This instrument offers an accurate index of the return of the progress of nerve regeneration and Would have been invaluable during the period of reconstruction in the selection of cases appropriate for nerve suture.

The indications for resection admit of little discussion: resection was recommended, when six months after injury the lesion was stationary, when there was evidence of a complete anatomical interruption, when there was an interstitial neuritis or a neuroma in continuity, without response to stimulation, especially if the neuromia was imperfectly encapsulated.

It was generallv conceded that one should resect sufficient of the damaged nerve to expose normal fasciculi, that is, above the seat of fibrosis. In some instances one had to stop short of conditions ideal for resection because the defect might be irreparable. A thickened nerve sheath in some instances was helpful, giving a better purchase to the suture under conditions of extreme tension. To relieve compression this nerve sheath, under these circumstances, was split longitudinally after approximation was effected.

Coaptation was effected with six to eight epineural silk sutures, and in many instances, when required to relieve tension at the line of suture, stay suture of catgut was employed. In tying the stay suture care was taken to avoid either crowding or a dead space between the peripheral and central segments. It has been claimed that an intraneural suture stimulates connective tissue formation, but the frequency with which the stay suture was used and the results obtained, must stamp this objection as of theoretical rather than practical import.

In the writer's experience it was only in an exceptional ease that a partial resection seemed indicated and usually in the lesions of the sciatic nerve, when either the internal popliteal or the external popliteal segment of the sciatic appeared intact. At this level one is dealing rather with two nerves than with one, although with a more intimate anatomical relationship than below the point of bifurcation. But partial resection as implying reconstruction of individual fasciculi is a much more delicate procedure. In Ney's experience this operation was not unusual.8 Each fasciculus was tested with the electrode; those that failed to respond were resected and coaptation effected with endoneural sutures.

It was universally acknowledged that a nerve bed free from scar tissue must be provided at the line of suture. Transposition to an adjacent intramuscular


plane or in the absence of this, to the plane of the subcutaneous tissue, was a common practice.

The protection of the line of suture with any foreign tissue was generally condemned. This applied to fascia lata, to veins or to Cargile membrane. Of the methods of bringing defects there seems to be no difference between those adopted by the surgeons of the United States and by those of other countries. Primary nerve stretching, nerve transposition. as with the median, ulnar and musculospiral nerves, and the two-stage operation were the selected methods, in the order mentioned. Implantation and reconstruction by anastomosis were regarded as illogical. The two-stage operation virtually eliminated the necessity for transplants and succeeded when the transplant failed. The writer employed resection of bone but once. It is too formidable to be considered as an acceptable procedure except. perhaps, in cases of musculospiral paralysis with an ununited fracture of the humerus.

Of the important factors in the technique of nerve suture the avoidance of nerve distortion has been given the most prominence. It has been assumed that a sensory fasciculus can not function as a motor and vice versa. But it is difficult to determine or to prove just what influence nerve distortion may have. In some cases after suture of the musculospiral nerve there was almost complete absence of formication in the radial nerve, predominantly sensory in function, and intense formication in the posterior interosseous. a predominantly motor nerved.8

There are divergent views as to the maintenance of a definite nerve topography from the plexus above the level at which the muscular branches are given off. Some observers, including Stoffel, 9 Marie,10 and more recently Kraus and Ingham, 11 maintained that the course of a motor fasciculus is straight from the point where the nerve has been made up to the point of offset of the fascicul as a branch. In opposition to this view may be cited the experiments of Dustin12 and McKinley,13 each independently finding vast plexuses in the constituent elements of a nerve trunk with at continuous exchange of fibers from the parent plexus above to the point where the peripheral branches are given off. These findings, true as they may be, do not negative the importance of avoiding distortion, since while there may be free interchange of fibers between the nerve supply of certain muscle groups, as for example between the fasciculi of the anterior tibial and peroneal nerves in the sciatic trunk the writer believes there is a more or less constant topography of the motor and sensory fasciculi in a given trunk, and after all, it is in the nerves of mixed function especially, such as the median and ulnar, that the maintainence of the proper axial relationship of the central and peripheral segments is important.


The following analysis was based in part on 3,129 peripheral nerve registers, representing the records of between 80 and 90 percent of the neurosurgical operations performed prior to the discharge of the patients from the Army or the transfer of their supervision to the United States Public Health Service. under the War Risk Insurance,5 and, subsequently,. to the Veterans' Bureau.6 Information concerning some of the peripheral nerve injuries included herein


was obtained by means of questionnaires issued through the Veterans' Bureau; from the published records of individual surgeons; from private communications from individual surgeons to the writer.

As stated above, between 80 and 90 percent of the operations for the repair of nerve injuries were performed while the men concerned still were under control of the Army, under conditions considered quite ideal. These operations almost without exception were performed by skilled neurosurgeons who had trained personnel as assistants. Until his discharge from hospital, whether operated upon or not, the patient had the advantage of such auxiliary and supplemental measures as were provided by well-equipped physiotherapy departments. From date of discharge (the majority by October, 1919) systematic treatment ceased and organized supervision for the most part was continued.

For purposes of determining what the results of operative treatment, were, only certain nerves were selected, because the injuries to them were in sufficiently large numbers to warrant conclusions concerning them. These nerves are shown in Table 10. Also, because combined lesions would cloud the issue, such lesions are omitted from consideration.

TABLE 10.- Certain operated cases, observed in the peripheral nerve centers

Of these 2,390 cases, 1,03S were operated upon and 1,305 unoperated. At first 23 to 30 percent was regarded as a conservative estimate of the proportion of cases in which operation would be required. Continued experience and observation proved this estimate too conservative as the figures just quoted indicate. 1.085 operated and 1,305 unoperated, in the ratio of 45 to .55 percent Since the cases under consideration do not include the operations performed subsequent to 1919, it is highly probable that such subsequent operations would equalize the proportion of operated and unoperated cases.

It should be understood that the examinations from which these end results have been compiled were not made by the officers stationed in the peripheral nerve centers. Because the patients were transferred to the cared the Veterans' Bureau, it was necessary to rely on the statements of physicians assigned to the various districts of that bureau through the country. It was impossible to secure through this source a detailed record by those familiar with the technique of peripheral nerve examinations and in the questionnaire distributed to the districts the officers were requested to record the results in terms of “good,” "mediocre," and "negative."



With the exception of the brachial plexus and the internal popliteal, the proportion of neurorrhaphies to neurolyses was fairly uniform, varying from 65 to 75 percent for the one and 22 to 34 percent for the other. In round figures about three-fourths of the cases required resection and suture and the remaining quarter the conservative procedure. In the earlier stages of the reconstruction period surgeons were inclined to adopt the more conservative policy, but with a larger experience and the observation of many failures following neurolysis, a more radical policy prevailed.

TABLE 11.- Proportion of neurorrhaphies to neurolyses                                                               

In estimating the merits of operations upon peripheral nerves the most convincing figures are those for the negative results, which represent failures. It is well known in the field of peripheral nerve surgery that perfect recoveries are only exceptionally obtained and the difficulty in defining sharply the grades of "good" and " mediocre" is manifest. Hence for purposes of contrast one should compare the cases which were outstanding failures and those in which there was a measure of success.

TABLE 12.- The percentage of good, mediocre, and negative results after neurolysis. a Indirect observation

That the results following neurolysis were not as good as was anticipated may be attributed to insufficient knowledge of the pathology before or at the time of the operation. External neurolysis, or neurolysis proper, was regardeid as appropriate in the compression syndromes when the essential lesion was the embedding of the nerve in perineural scar tissue. Assuming that the fasciculi were intact, that there were no intrafascicular adhesions, no interstitial cicatrices. one wonders how much neurolysis proper may have influenced the results. Would there not have been spontaneous regeneration in many of the cases whether neurolysis had been performed or not?


Endoneurolysis is an entirely different procedure and was used to advantage in a certain number of cases when there was evidence of an intraneural lesion. A free incision was made through the sheath when the latter was fibrosed; if the fasciculi were bound together with adhesions these were separated. But if the nerve was the seat of a frank interstitial neuritis, resection was considered essential.

The compilation of the end results in neurorrhaphy (Table 13) were based on a series of 400 cases. As previously stated, the examinations, upon which these figures are based, were not the personal examinations of the writer or of the operator, but were made by a number of physicians, not necessarily neurosurgeons, employed in the Veterans' Bureau. Hence the results must be considered approximate.

TABLE 13.- The percentage of good, mediocre, and negative results in motor function after neurorrhaphy b Indirect observation

Compared with the statistics from other sources, the end results, again measured chiefly in terms of failures, do not vary very widely, except. in the case of the external popliteal nerve; the percentage of negative results is unaccountably high.

TABLE 14.- The percentage of good, mediocre, and negative results in motor function in the total series of operated cases, including neurorrhaphy and neurolysis. c Indirect observation

In addition to these 400 cases of neurorrhaphies and 119 neurolyses, considered in Tables 11 to 14, a table of 497 operated cases (Table 15) including 132 neurolyses. 350 neurorrhaphies and 14 transplants, is given for comparison. This table is of greater value since the examinations were made mostly by an individual surgeon experienced in this particular field.


TABLE 15.- Percentage of end results of 497 operations, including 132 neurolyses, 350 neurorrhapies, and 14 transplants a

The best results were obtained in the internal popliteal and median nerves with a failure in only 8 and 14 per cent respectively and "good" or "mediocre"results in 92 and 86 percent respectively. The largest percentage of failures. as expected, was in the operations on the musculospiral nerve with failure in almost 50 per cent. The ulnar and sciatic nerves occupy an intermediate position with results not far apart, approximately 30 percent failures and 70 percent "good" or "mediocre."

The results en masse in the 470 operations yielded 34 percent "good" results, 36 percent "mediocre" and 26 percent "failures." In other words in any large series of cases we may anticipate good or mediocre return of function results in two-thirds of cases, negative results in one-third.

TABLE 16.- Percentage representing Tables 14 and 15 combined

Comparison of statistics in peripheral nerve surgery is a questionable practice. The matter of interpretation of the words "good" and "mediocre is a question in point. What one individual might record "good," the other might consider "mediocre." Therefore allowances must be made for the source of information-whether obtained by direct personal examination of the operator or indirectly through a physician unskilled in these very technical examinations, or perhaps from the patient himself.

No matter with what infinite care and with what nicety of approximation a nerve suture may be effected the percentage of successes and failures will depend upon the regenerating forces of nature. One can never tell what changes may have taken place in the spinal cells. It is with relation to the activity of the spinal cells that the time element becomes an influential factor. This factor, together with the degree of degeneration of the peripheral segment and the degree of atrophy or fibrosis of the muscles involved, must in many instances determine the end result.



In 17 experiments conducted by Huber14 a defect of 3 cm. was successfully bridged by an autotransplant, and in 6 experiments a homotransplant was used with evidence which justified the indorsement. He was successful not only with fresh homotransplant but also with those stored in 50 percent alcohol for 40 days. These brilliant results in the experimental laboratory in the use if the transplant, as a means of repairing defects, are in striking contrast, to the reports from the peripheral nerve centers.

In approximately 1,414 operations upon the peripheral nerves in the Army hospitals there were in the neighborhood of 60 transplants used to repair defects. Of this number the writer has been unable to find the record of any "successful" result, except in a few isolated instances. In one "very marked improvement" is recorded after the use of three strands of a cutaneous nerve to repair a defect 6 cm. long in the musculospiral nerve; in another,"considerable improvement after one year" is recorded in a defect of 4 cm. in the external popliteal nerve, to repair which strands of a cutaneous nerve were used. In one instance the employment of an auto transplant to repair a defect in the ulnar nerve at the wrist was followed by definite contraction of the zone of anesthesia. One neurosurgeon reported that in 14 attempts there was but case of transplant in the median nerve where, after four years, the patient's only sensation was a sense of tingling in the median distribution. That there is little need for the use of transplants to repair defects may be gathered from the experience of individual operators. In one series of 196 operations an auto-transplant was used only three times. In another of 570 operations an auto-transplant was used six times and homotransplant eight times. Considering the total number of cases and the results as recorded, the employment of the transplant either " auto" or " homo" as a practical method of bridging defects in peripheral nerves has proven a dismal failure in the hands of the surgeons of our country. The results of nerve stretching in a two-stage operation for the correction of large defects, even when a nerve is sutured under great tension, greatly surpass these obtained from the use of the transplant.

Why the results in the experimental laboratory can not be reproduced in human surgery has never been explained. To be sure, there are physical factors in the pathology of peripheral nerve lesions of the human that are wanting in the experimental animal, and what is of no small moment, the length of the-raft employed in the experimental laboratory is only one-fourth or one-half that required to repair the defect in the resections of extensive peripheral nerve lesions. Just as the transplant has proved successful in the laboratory, so has lateral implantation of the peripheral and central segments into an adjacent nerve, but so far as the writer is aware, neither the nerve flap operation or nerve crossing or implantation has been applied successfully in the reconstruction of peripheral nerve injuries. Both operations seem illogical, and neither has found favor with those who, in dealing with hundreds of cases, have acquired an intimate knowledge of the problems involved.


(1) Based on sick and wounded reports made to the Surgeon General.
(2) Annual Report of the Surgeon General, U. S. Army, 1919, ii, 1096.
(3) Letter from the Surgeon General to Maj. George Muller, M. C., January 31, 1919. Subject: Peripheral Nerve Commission. On file, Record Room, S. G. O., 024.14 (Surgery of the Head).
(4) Clinical Records, entitled "Peripheral Nerve Register." On file, Record Room, S. G. O., 700.6-1.
(5) Act of Congress, approved June 27, 1918; also, Act of Congress, approved March 3, 1919.
(6) Public Act No. 47, 67th Congress, August 9, 1921.
(7) Sachs, Ernest, and Malone, Julian Y.: A More Accurate Clinical Method of Diagnosis of Peripheral Nerve Lesions and of Determining the Recovery of a Degenerated Nerve. Archives of Neurology and Psychiatry, Chicago, 1922, vii, No. 1, 58.
(8) Ney, Karl Winfield: The Indications for Surgical Intervention in Peripheral Nerve Injuries. Journal of the American Medical Association, Chicago, November 8, 1919. lxxiii, 1427.
(9) Stoffel, A.: Die moderne Chirurgie der peripheren Nerven. Medizinische Klinik, Berlin, August 31, 1913, ix, 1401. Also: Vulpius, Oskar and Stoffel, Adolf: Orthopädische Operationslehre. Ferdinand Enke, Stuttgart, 1911.
(10) Marie, Pierre: Les localisations motrices dans les nerfs periphériques. Bulletin de l'académie de médecine, Paris, December 28, 1915, 3 s. lxxiv, 798.
(11) Kraus, Walter M., and Ingham, Samuel D.: Peripheral Nerve Topography. Archives of Neurology and Psychiatry, Chicago, 1922, iv, No. 4, 259.
(12) Dustin, A. P.: Le service de neurologie a l'ambulance "Ocean." Travaux de l'ambulance Ocean. Masson et Cie, Paris, July, 1918, ii, 135.
(13) McKinley, J. C.: The Intraneural Plexus of Fasciculi and Fibers in the Sciatic Nerve. Archives of Neurology and Psychiatry, Chicago, October, 1921, vi, 377.
(14) Huber, G. Carl: Repair of Peripheral Nerve Injuries. Surgery, Gynecology and Obstetrics, Chicago, 1920, xxx, No. 5, 464.