Shoulder:Nerve Problems Related to the Shoulder

Bullet Points

This subsection does not exist. You can ask for it to be created, but consider checking the search results below to see whether the topic is already covered.

Introduction

There is potential for neurologic injury during most open or arthroscopic surgical procedures as the brachial plexus and its terminal nerve branches are in close proximity to the glenohumeral joint. The positioning of retractors, arm positioning, distorted anatomy in multiple operated patients or patients with fracture sequelae, biomechanical changes after prosthesis increase the potential risks for neurological lesions during shoulder surgeries. Moreover, numerous entrapment syndromes have been described. Consequently, neurological lesions around the shoulder are frequently reported. They must be recognized and treated properly in order to guaranty to the patient the best results have.

Anatomical Considerations

Peripheral nerves are composed of axons, myelin sheaths, connective tissue and blood vessels. A disturbance of nerve function can occur during abnormalities of each of these structures. Each nerve is anatomically divided into individual fascicles, themselves being surrounded by dense connective tissue: the epineurium. The perineurium lies deep in the epineurium. It corresponds to an extension of the pia mater and the arachnoid, functioning as a protective membrane. Endoneurium is an interstitial connective tissue of axons. The internal structure of an axon consists of linear microtubules, essential for axoplasmic flow, and microfilaments. Interruption of this flow causes dysfunction and possibly axonal death with distal degeneration.

The peripheral nerve is a well vascularized structure with two integrated but independent systems: microvascular systems. The vessels of the endoneurium are normally impermeable, but following trauma, their permeability may increase, resulting in an endoneural edema that may interfere with nerve function, reflecting an irreversible injury.

The brachial plexus is formed from the coalescence of the ventral rami of the lower four cervical (C5 through C8) and first thoracic (T1) spinal nerves as they emerge from their respective neural foraminae. It is responsible for most of the innervation to the upper limb. The C5 and C6 nerves merge to form the superior trunk, and the C8 and T1 nerves merge to form the inferior trunk. The C7 nerve continues laterally as the middle trunk.

Classifications

Depending of anatomical nerve damage

There are two main classifications in acute: the one of Seddon and the one of Sunderland. Both are based on the different structures of the affected nerve and are very well correlated. Seddon's classification uses the terms neurapraxis, axonotesis, and neurotesis. Sunderland's classification uses five degrees:

1. Neurapraxis or stage 1 refers to a conduction block, indicating that the nerve and axon are intact but do not lead to action potentials;
2. Axonotesis includes an axonal rupture. The surrounding Schwann cells are preserved in stage 2;
3. Stage 3 has a simultaneous rupture of Schwann cells adjacent to the endoneurium;
4. In stage 4, the rupture of the fascicules, including the perineural sheath is total.
5. Neurotesis or stage 5 refers to a complete rupture of the nerve and results in permanent loss of nerve function unless surgery is performed.

The type of nerve damage significantly influences future prognosis, with ultimately more or less nerve repair depending on the situation. Functional nerve recovery is possible with neurapraxis and axonotesis.

Sunderland and Seddon Classifications

Sunderland classification	Seddon classification	Lesion	Clinical signs	Recovery
Sunderland I	Neuropraxia	Intrafascicular edema, segmentary demyelisation	Paresthesia, paresia (partial or complete)	Complete (days to 3 months)
Sunderland II	Axonotmesis	Axonal lesion, preservation of the connective tissue framework of the nerve (the encapsulating tissue, the epineurium and perineurium	Paresthesia, paresia (partial or complete)	Generally complete (1-6 months
Sunderland III	Axonotmesis	Lesion of the endoneurium, epineurium and perineurium remain intact	Paresthesia, dysesthesia, paresia (partial or complete)	Partial (12-24 months)
Sunderland IV	Axonotmesis	Only the epineurium remain intact	Hypoesthesia, dysesthesia, complete paralysis	None without surgical repair
Sunderland V	Neurotmesis	Disruption of the entire nerve fiber	Anesthesia, complete paralysis	None without surgical repair

Wallerian degeneration (degeneration of the part of the cell that does not contain a nucleus, the axonal extension in the case of the nerve cell) occurs for all lesions of axonotesis or neurotesis.

Depending on the type of mechanism of nerve damage

Nerve section

It can be partial or complete and more or less clear (often lesion by knife, firearm ...)

Lesion by stretching

As in the case of displaced fracture or reverse shoulder arthroplasty for example. The degree of neurological deficit depends on the degree of stretching. During traction, the perineurium elongates, and the previously undulating axon becomes stretched. Intraneural pressure exacerbates lesions of a vasculo-nervous bundle. This type of injury can lead to continuous neuroma. If the lengthening continues (>8% of total length of the nerve), the axon stretches all the more and begins to present tears, likewise for the perineurium and epineurium. This type of injury can happen in acute or chronic way.

Nerve compression

May be intrinsic (tumors) or extrinsic (surrounding or external structures such as retractor). The mechanism results either in a direct compression of the nerve, or indirectly. In the first case, there is damage to the myelin sheath or the axon itself, thus limiting nerve conduction. In the second situation, the lesion is made by vascular compression, resulting in an endoneural edema that will disrupt the nerve function directly by modifying the axonal flow or indirectly by increasing the pressure of the endoneural fluid (like a syndrome of the lodges). These disturbances (edema and microcirculation disorder) associated with chronic irritation leads to fibrosis and other intra and extra-neural scars.

Nerve friction

Occurs when there is dynamic or intermittent compression resulting in friction of the nerve against the ancillary structures. This irritation stimulates the fibroblasts in the epineurium, resulting in the compression of fascicles within the epineurium itself.

Pressure modification in anatomical tunnels

This situation is found in the ulnar tunnel for example, where during the flexion of the elbow, there is an increase in ulnar nerve pressure.

Nerve constriction

This mechanism is rare and the pathophysiology is unclear. However, this type of lesion is close to a complete section (Sunderland stage V).

Sunderland's classification describes well anatomical lesions but does not take into account certain preoperative factors that may influence recovery (such as loss of substance, local ischemia, burns, infections…). Similarly, the age of the patient, the smoking or the delay significantly reduce chance of recovery. A specific scale to predict final prognosis of a nerve injury have been developed. Prognostic scale takes into account the anatomical classification of Sunderland and the most important factors affecting the prognosis of nerve recovery. This scale consists of 15 points. The more points, the poorer the prognosis. A nerve injury scoring 1 point has an excellent prognosis, whereas a lesion with 15 points has a very poor prognosis.

Nerve Injury Severity Scale

Nerve Injury
Sunderland I	1
Sunderland II	2
Sunderland III	3
Sunderland IV	4
Sunderland V/VI	5
Motor nerve grafting	6
Type of Injury
No associated injuries	0
Crush, avulsion injury	1
Associated injuries(fracture, ischemia)	2
Delay of motor nerve repair (months)
< 6	0
>6 delay < 12	1
>12	2
Level of motor nerve injury
Distal injury	0
Proximal injury	1
Age of the Patient (years)
Age <20	0
>20 age <50	1
Age >50	2
Perioperative Smoking
No	0
Yes	1
Management in Microsurgical Unit
Yes	0
No	1

There are several mechanisms of recovery following motor nerve damage: the resolution of conduction block during neurapraxis, muscle fiber hypertrophy during neuropraxis and axonotesis, intra-muscular collateral reinnervation during partial axonotesis and axonal regeneration from the site of lesion during partial and complete axonotesis and neurotesis.

Recovery of strenght over time after nerve damage

The following explanation may allow to understand the poor functional recovery occasionally observed: in axonotesis, axonal continuity is lost but the structures supporting the nerve are preserved. Axonal degeneration occurs on the distal portion of the axon from the site of the lesion. However, the preservation of the endoneural sheath allows the axons to direct their regeneration pedicles towards the originally innervated organ. This type of lesion should have a relatively good prognosis. However, the distance between the site of the lesion and the innervated organ vary. As the axon only regrows distally at an average speed of 1 mm per day, the motor plate and muscle fibers begin to atrophy before an axon reaches them. Finally, Schwann cells in the basement membrane degenerate after prolonged periods and the chances of an axon finding its way diminish for lesions located far from the target organ. Time limits thus the potential for reinnervation. If the axon does not reach the motor plaque within two years, the degree of atrophy and scarring of the target muscle becomes so important that restoration of motor function is unlikely.. Moreover, if the motor plate is affected, a sufficient number of axons is required for the muscles to be functional. So there is not always parallelism between electrical and clinical recovery. Indeed, it seems that for an electric reinnervation to be effective, it must be early; otherwise, recovery is still insufficient to restore a function.

Clinical Examination

The sensory examination consists of light touch, deep touch, pain, temperature, position sense, and vibration evaluation. However, in clinic, light touch is the modality most frequently used. There are only three cervical reflexes: the biceps (C5), the brachioradialis (C6), and the triceps (C7). In the upper extremity, the best long-tract sign is the Hoffmann reflex, which is analogous to a Babinski reflex in the lower extremity. Muscle strength is graded from 0 (no activity) to 5 (normal). Clinical differentiation between neurapraxis, axonotesis and neurotesis is difficult, especially at the early stage.

Brachial Plexitis (Parsonage-Turner Syndrome)

Brachial plexitis is a condition characterized by acute onset of shoulder pain followed by weakness and/or sensory loss of the shoulder and/or upper extremity. Dreschfeld first described it in 1887. The most important contribution to brachial plexitis was the report of 136 patients by Parsonage and Turner in 1948. The idiopathic form is associated with recent viral infection and vaccination. The hereditary form is an autosomal dominant recurrent brachial plexitis. The incidence is estimated to be approximately 1.64 cases per 100,000 person-years. The disease commonly affects males during their third and seventh decades. The most common initial symptom is acute onset of intense pain around the shoulder followed by weakness. Winged scapula may be observed in 20% of cases (Figure). Thirty percent of the cases are bilateral. Electromyography is a key clue toward the diagnosis. MRI on T2 sequences reveals edema secondary to nerve demyelination. The treatment of brachial plexitis is conservative, based on analgesia and physiotherapeutic rehabilitation.

Right winging scapula due to Parsonage-Turner syndrome.

Thoracic Outlet Syndrome

The thoracic outlet syndrome is caused by compression of the nerves of the brachial plexus or the vascular bundle in the arm between the first rib, the middle scalene muscle, and the clavicle. In the neurogenic form, the brachial plexus is compressed beneath the scalene muscles or in the costoclavicular space. This syndrome is difficult to diagnose accurately because there is no specific history or physical examination finding. Pain, atrophy, weakness or paresthesias are commonly found.

What would Codman have thought about this?

In 1929 when I had already spent a year on this book, I found, by mere chance, that my neighbor, Dr. Stevens, had for a long time made a hobby of studying the brachial plexus. I had often talked with him about other injuries of the shoulder, subjects on which he had made some important contributions, but he had never mentioned that he had been so much occupied with studying lesions of the brachial plexus. I found that he had never written anything on the subject because, after all the years of study he had devoted to it, he felt that little was really known about these lesions and that his personal views would interest very few readers. I asked him if he would not review his notes and write a chapter for my book. Although he said that he could add very little to the ordinary textbook knowledge about the subject, I assured him that if he did give me the benefit of what study he had made, it would at least bring the subject up to date and show that it was open for still further investigation. I at length persuaded him to try. He had at that time dissected ninety-two plexuses through the courtesy of the Tufts Medical School, and since then, through the kindness of the anatomical department of the Harvard Medical School, he has been able to dissect a good many more, but owing to his illness during part of this time, he was only able to analyze sixteen of these later dissections for this chapter. However, these were done very carefully, with especial reference to confirming his previous observations.

On March 24,1932, Dr. Stevens died suddenly from heart disease, leaving his work in preliminary manuscript form for me to use. The following chapter has been constructed from his manuscript, using largely his own words, but modified to some extent because I was obliged to abbreviate parts of it, and in other parts to add a few paragraphs to clarify his meaning. The substance of the chapter is his and he deserves the credit for all the originality which it contains. I have used the portions of his manuscript which seemed to me original and enlightening, although in some cases I have not actually verified his observations or presumed to criticize his conclusions. For instance, I have made no dissections to verify the new nerve filament which he describes, nor can I take any stand in regard to his views about prefixed and postfixed plexuses. Neither could I review and tabulate the 710 cases which he has collected and analyzed. I have merely set down his conclusions and can only say that they seem to me valid.

Dr. Stevens was one of the most independent-minded men I have ever known. He practiced surgery for many years in Boston, but was never associated with any of the large clinics. Nevertheless, he was a true student of medicine, and took the greatest interest in his own individual patients. Any case which he could not thoroughly understand led him to study the subjects involved most carefully, not only by searching the literature, but by experimental work in the machine shop or in the dissecting room. He had little respect for modern authority and always tried to trace his subject to the original writers who first described it. For these pioneers in medicine he had the greatest respect and admiration, and a corresponding scorn for plagiarists. As the years went by, his habit of giving intensive study to each obscure case which occurred in his practice, gave him a remarkable all-around knowledge of surgery and its problems. Few men in their fifties and sixties have the industry to dissect out the branches of the brachial plexus to satisfy themselves about mooted points which they have unearthed through finding differences of opinion among famous authors. He seldom spoke at medical meetings for he was of a retiring disposition, although on superficial acquaintance he seemed assertive. I have never heard him "read a paper", but he published several papers on fractures, which to my mind show an unusually clear grasp of the subjects. It was very characteristic of him to have done all this work and more on the plexus and to have made no attempt to publish it. He enjoyed satisfying his own curiosity, but he hated the details of publication. I feel that I was very fortunate in being able to rescue from oblivion some of the thoughts suggested by his investigations. At any rate, I believe that the surgeon who is willing to give proper study to this chapter will agree with me on finishing it, that he has a clearer and better idea of brachial plexus injuries than he ever had before. One feels that the subject has been very carefully studied by the author, and that his views about the essential mechanics governing these injuries have been thoughtfully presented, and are not likely to be challenged.

Causes of brachial plexus paralysis

Traction

a. Birth cases.

b. Blows on shoulder or neck depressing shoulder, or fulls stretching head away from shoulder.

c. Falls when the arm catches, suspending the person at least momentarily, or where arm is twisted.

d. Being lifted by the arm as over a pulley, or dragged forcibly by the arm as in case of children jerked by parents leading them by the hand.

e. Accompanying dislocations or the reduction of dislocations.

f. Prolonged holding of arm in abduction and external rotation, as in operation.

g. Slipping while carrying a weight on shoulder.

h. Sudden movements, as in effort made in trying to lift weights.

Pressure

a. Dislocation, where shoulder is out for a long time.

b. Exuberant callus, especially of the clavicle.

c. Saturday night paralysis, usually radial, terminal.

d. Crutch paralysis, terminal.

e. New growth, involving roots, trunks or terminal branches, secondarily, either from bony, cartilaginous or soft tissue pressure (benign or malignant).

f. Adventitious rib or bands.

g. Strait-jacket; several cases have been reported, but always where the patient freed the arm so as to cause constriction.

h. Following aneurism of the subclavian or axillary artery.

i. Esmarch bandage or tourniquet.

j. Degenerative bone lesions of the cervical spine, especially tuberculosis, causing pressure on roots.

k. Fractures; especially of vertebras, by direct pressure.

Direct Injuries

a. Cuts or wounds.

b. Fractures or dislocations causing direct injury.

c. Injection of the plexus by local anaesthetics.

d. Electric shock.

Toxic

a. Following serum injections.

b. Infectious diseases, especially pneumonia.

c. Following anaesthesia, chloroform especially (these are hard to separate from possible trauma from mal-position during the anaesthesia, but several cases have been reported where no mal-position could have occurred).

d. Poisons—as lead—usually peripheral neuritis.

Spinal Cord Lesions

a. Infection.

b. Degenerative processes.

c. Tumors.

d. Direct injuries to the cord itself.

e. Hemorrhage within vertebral canal.

Central

Monoplegia. Rare; never complete; spastic, not flaccid.

This list covers most of the causes of injury, but cannot indicate the relative frequency in which they occur; no such statistics have ever been collected. As a matter of fact, in a very large majority the injury is of the traction type. In a smaller number it may be of the pressure type. The lesions caused by the remaining five types are either too rare or too obvious to be discussed here. Most of the common injuries have to do with temporary, partial or complete displacement of the head of the humerus. The plexus may be injured by the same force acting concomitantly directly on the plexus, or by the prying effect of the displaced head on the cords themselves. If the head of the humerus remains out of position for some time, direct pressure on the plexus itself may be a factor. As indicated in the diagram below, brachial plexus injuries may be classified for purposes of study in several ways besides the etio-logic grouping given above.

FIGURE 63. DIAGRAM OF NEUROVASCULAR CORD

The primary division into supra- and infra-clavicular lesions is important, not only because in the literature of the subject one finds endless discussion as to which group is the more common, but because when surgical exploration has been decided upon, one must choose between the upper or lower field for the incision. The secondary divisions obviously cannot be made strictly in subdivisions of the primary division, for a small part of the trunk lies above, and a large part below the clavicle. All root lesions are necessarily above the clavicle, and all lesions of the great terminal branches, including the axillary, are below the clavicle. However, some of the smaller but highly important upper terminal branches, usually spoken of as "the root collaterals," such as the dorsalis scapulae, suprascapular, phrenic and long thoracic, are above the clavicle since they arise from the roots, or from the fasciculi not far from the roots. In the text which follows, we shall frequently allude to that part of the plexus between the junction of the roots and the origins of the great terminal branches as the "trunk." This trunk when dissected is the "plexiform part." Since in most of its extent the "plexi-form part" forms a compact bundle about the axillary artery, we also speak of it as the "neurovascular cord" or "integrated cord." This fascia-bound structure is a very real anatomic entity, although somewhat vague in all its limits, since each structure that enters or leaves it contributes or carries away accompanying strands of fascia. It is a single cord closely integrated by dense fascial investment from just beyond the interscalene segments to well below the shoulder, a single cord which flares out at the base into five roots of varying sizes, and, far below the apex of the axilla, flares out again into its terminal branches.

Root paralyses. The Erb-Duchenne, or upper type of brachial plexus palsy, is the most common form of paralysis from root injury. It was described in detail with only minor differences by both Duchenne (1872) and Erb (1874). It is accepted as a paralysis of the supra- and infra-spinatus, the deltoid, the biceps, the brachialis and the brachioradialis. If the lesion is far back on the roots, we are told that the paralysis usually includes the levator anguli scapula, the rhomboids, and sometimes the serratus anterior and the portion of the diaphragm supplied by the phrenic nerve. If these muscles are paralyzed it clinches the root diagnosis. If they are not, it does not exclude it, for anomalies occasionally exist. Accurate statements as to the condition of the muscles are too seldom included in most of the reports of examinations of cases of brachial paralysis. Many cases of deltoid, biceps, brachialis and coraco-brachialis paralysis have been reported as Erb-Duchenne, but inclusion of the supra- and infra-spinatus and the brachioradialis is necessary for a diagnosis of this type. The coraco-brachialis, the supinator, and, as Harris pointed out years ago, the extensor carpi radialis longus and brevis are sometimes also included in this type, as will be explained later. The pronator teres may also be involved, for it frequently is supplied from the 6th cervical nerve. When the posterior part of the deltoid, which is also probably supplied by the 6C, and the pronator teres are involved, it raises the question of inclusion of the 6th at least. Probably the innervation of the pronator teres may sometimes come from the 7C. That of the coraco-brachialis and extensor carpi radialis longus and brevis may also more often come from the 7C than from the 6C. Harris called attention to the fact that the inclusion of the 6C in a root lesion usually added nothing to the extent of the paralysis caused by a 5C involvement, and Sherren and other observers have agreed with him. So that the 6C is to be regarded as a sort of makeshift root, its inclusion adding little to the severity of the paralysis. The clinical term "Erb-Duchenne paralysis" has come to be used very loosely. It no longer defines, as it should, a root paralysis of the 5C and 6C, but is extended to include cases which involve the 7C. The involvement of the 7C should place the case in the complex type. The term is also applied thoughtlessly to peripheral lesions of the axillary nerve, and even to cases of lead palsy in which the axillary is involved. Moreover, the term Aran-Duchenne paralysis is sometimes confused with it, and is erroneously used. This term indicates an entirely different pathologic condition, due to a degenerative lesion of the anterior column cells. It usually first affects the small muscles of the hand and is slowly progressive, and finally at least is bilateral. The careless use of these terms, without definite statements as to exactly which muscles are paralyzed in each reported case, renders many articles worthless for purposes of detailed study. It is usually agreed that there is remarkably little sensory disturbance with the Erb-Duchenne type, and that this is confined to the axillary distribution on the external or postero-external surface of the shoulder; but all observers are by no means in accord with this, especially Rendu. Andre Thomas's two cases, of cut 5C and 6C, and of cut 6C root alone, both showed a well-marked sensory disturbance, not confined to the axillary distribution.

The lower or Dejerine-Klwnpke type, due always to an involvement of the 8C and ID roots, is a combined paralysis of the flexors and extensors of the forearm with integrity of the brachioradialis, supinator, pronator teres and the extensor carpi radialis longus and brevis. The intrinsic hand muscles are paralyzed and the upper arm extensors (triceps) are partially involved. In this type sensory disturbances are much more profound. Trophic and vasomotor symptoms are sometimes pronounced, although some observers believe that in a pure brachial plexus lesion, vasomotor and trophic symptoms are not pronounced. As few such injuries are pure, this is a fine distinction. Vasomotor and trophic fibers come via the autonomic to their somatic nerve distribution. The syndrome of Poirier, or as the French call it today, the syndrome of Claude Bernard-Horner, is present when the ID is involved far back on the roots, and is caused by involvement of the rami com-municantes which pass to the stellate ganglion. It is shown by miosis, enophthalmos and ptosis on the side of the lesion, with preservation of the light reflex and accommodation; there is normal tension or slight hypotension and normal vision, but abolition of dilatation of pupil to cocaine. Miss Klumpke's* contribution to the subject of brachial plexus injuries has given her name to this lower root type. Her report included three cases of her own which she followed in the Clinics of Vulpian and Lucas De Championiere, but only one was of this lower type. None of her own cases was proven by operation. The rest of her eighteen cases had been reported by other men and included the case of Flaubert. Her real contribution was her experimental work on animals, by which she proved that the syndrome of Poirier or Claude Bernard-Horner followed a section of the first dorsal root back of the ramus communicans to the stellate ganglion, and that it occurred only after section of the ID or of the ramus itself.

• Miss Augusta Klumpke (1859-1927) was the eldest of four brilliant American sisters who were born in California and later educated in Switzerland. She studied medicine in Paris and achieved the distinction of being the first woman interne to serve in the Paris hospitals. In 1885, soon after her graduation, she published an article on paralyses due to injuries of the lower roots of the brachial plexus, a type which has since been designated either as "Klumpke paralysis," or as "Dejerine-Klumpke paralysis." This confusion has arisen because five years after the publication of this paper she married Professor D6jerine, a neurologist and also a pupil of Vulpian's. They were later joint authors of numerous books and papers considered among the foremost landmarks in the progress of modern neurology. The three younger sisters also became distinguished in their chosen fields, of astronomy, music, and painting. The youngest, Miss Anna Klumpke, is widely known as an artist and as the friend and biographer of Rosa Bonheur.—E. A. C.

An intermediate type of brachial plexus paralysis, due to involvement of the 7C root alone, is practically non-existent, except as an accompaniment of either the Erb-Duchenne or the Dejerine-Klumpke types, when it would be classified as " Complex." The pure 7C paralysis would be a partial paralysis of the extensors of the arm, forearm and wrist, including the abductor longus pollicis and the extensors of the thumb. As the coraco-brachialis and extensor carpi radialis longus and brevis and the pronator teres are supplied by the 7C at times, these would therefore sometimes be included in the paralysis, but as these last come also at times from the higher roots, they might not be involved. A knowledge of the condition of the root and trunk collaterals from the 7C and the intermediate fasciculus, would not be of as great help in diagnosis as in either of the other types. The lower part of the serratus anterior might show a paresis, but as it receives branches from the 5C and 6C, it would not be entirely paralyzed. The same applies to the subscapularis and the latissimus dorsi, which, while receiving fibers from the 7C, also receives at least as great an innervation from other roots.

Complex types. As a complication of the upper type, the inclusion of the 7C adds to the usual group of muscles the following: partial paralysis of the extensors of the arm, forearm and wrist, including the extensors of the thumb; a more profound involvement of the trunk collaterals to the latissimus dorsi and teres major; the entire paralysis of the subscapularis, the coraco-brachialis, and the epicondylar muscles. As an accompaniment of the Dejerine-Klumpke type, the inclusion of 7C is indicated by the addition of absolute paralysis of the extensors of arm, forearm and thumb to the other symptoms of a lower type paralysis. In this type the subscapularis paralysis is never entire, but the latissimus dorsi is likely to be entirely paralyzed. The pectorals are seldom entirely paralyzed except in a complete lesion of all roots. The upper part of the pectoralis major, i.e., the clavicular portion, is included in an upper type, and the lower part is involved in a lower type together with the pectoralis minor.

The complete type. In very rare cases where the arm has almost been pulled off, all, or nearly all, of the roots of the plexus may be injured. In the period from 1870 to 1890, there was much discussion, particularly among the French, as to the exact location of root injuries, i.e., whether the rupture occurred within the spinal canal and involved the cord (avulsion de la moelle epiniere), or in the bony gutters formed by the transverse processes, or between the point of exit from these gutters and the point where the roots join to form the fasciculi. The term "radiculo-medullaire" was also used in an almost synonymous sense with avulsion. It is probable that today the French neurologists may have dropped these terms or attached other shades of meaning to them than those which we have interpreted. However, we must present these terms because they have been so much used, although the writer is satisfied from his own observations on the cadaver that nearly all injuries must occur entirely outside of even the bony gutters, because careful dissection shows that the roots are always snubbed at the transverse processes, i.e., attached to them by fascial connections. The reasons for believing that this snubbing will usually prevent injury to the roots within the canal, or even within the bony gutters, will be given later. Some authors (e.g., Shallow) have found at operation cysts containing cerebrospinal fluid close to the transverse processes, and have concluded that this was evidence of an evulsed root. This does not seem to me sound, because tubular processes from the dura not infrequently extend beyond the transverse processes and would be ruptured with the snubbing. (Fig. 63.)

Gross anatomy of the intact plexus. Before we take up in detail the study of the distribution of the nerves to the various muscles of the arm, it is necessary to consider the plexus as a gross unit, for we must always remember that in the body it has not been dissected and does not lie in the form of a diagram. In the last sixteen particularly careful observations I found that each root of the brachial plexus was held firmly at the transverse processes and in the gutters of bone by invaginations of the prevertebral fascia., In the words of the mechanical engineer, each root is snubbed at the transverse processes. There it is firmly fixed. If it were not, and there was nothing to take the ordinary strain away from the spinal cord itself, a man might be paralyzed at any time by a strong pull upon either the arm or shoulder. Below the clavicle, the cords of the plexus are firmly bound together and also to the surrounding structures, including the clavicle and the coracoid, by layers of the clavo-pectoral fascia. The three main cords join the subclavian artery as soon as it appears above the first rib, and remain integral with it in its course through the axilla. The clavo-pectoral fascia comes down from the back of the fascia which surrounds the subclavius muscle, forming the costo-coracoid membrane. This membrane goes to the coracoid and joins the axillary fascia below. The clavo-pectoral fascia leaves the back of the costocoracoid membrane, passes between and invests each branch of the brachial plexus, surrounds the axillary artery and is continuous above with the prevertebral and below with the deep fascia over the first rib. Therefore, the integrated neurovascular cord is snubbed again, as well as at the transverse processes, by these fascial investments between the clavicle, the coracoid and the first rib; that is, in the apex of the axilla. I should like to acce"nt the facts in the preceding paragraphs because I believe that these points of attachment help to explain the different types of lesions resulting from a blow or stress above the clavicle, and also those resulting from stresses which come through the arm from below, either as straight tension, or as lateral stresses resulting from dislocation. I have been interested in the mechanics of the shoulder for many years, and have studied on the cadaver the probable relation of the plexus to the displaced bones in cases of fracture and of dislocation, and the conclusion has been reached that the explanations found in the literature of the mechanism of the involvement of the plexus in these injuries are not adequate. I shall discuss the theories which have been offered to explain the mechanics of the more common lesions of the plexus from trauma, and then offer one which I believe shows why the plexus is either injured at its roots beyond their exit from the spine and before they join with others to form the plexus, or in its terminal branches, and never (except by cuts or wounds) in the plexiform part, which, nevertheless, is usually involved secondarily by hemorrhage and exudate between the nerve bundles. Most observers, up to the time of Horsley, believed in a local injury to the brachial plexus, both in birth and adult cases. Erb believed that the injury was due to clavicular compression, but this idea was gradually destroyed by Gerdy, Fieux, and Horsley. The idea of root lesions originated with Secretan, and thereafter there was a division of these cases into root, trunk and terminal lesions. To Gerdy is given credit by the French of the traction theory, that is, straight pull as the cause of these paralyses, although Malgaigne knew it long before this, for he had paralyzed two cases himself in reducing dislocations by a direct pull of 200 to 250 kilos (more than 500 pounds). Flaubert must have known it, since he had thus caused paralysis not only in his celebrated case of 1827, but also in two others, by direct traction. Various experiments have been made, and a great amount of time has been spent on the mechanics of these lesions. The case of Flaubert was the starting point of the avulsion theory, that is, the separating or pulling out of the roots from the spinal marrow itself. This, we contend, is rare except in very violent injuries. To Horsley, in England, is credited the idea that the lesions are caused by injuries which tend to spread the head and shoulder apart and thus to stretch the plexus. Today this is generally accepted as the most likely theory of production. Horsley's experiments are not given in detail, but in the Practitioner, of London, he said that by dropping cadavers on head and shoulder, he had been able to break the plexus.

Duval and Guillain, in 1898, made an extensive study of the plexus anatomically, carefully estimating the different angles formed by the individual cords from their origin at the spinal marrow to their emergence under the transverse processes of the vertebrae and across the neck to the arm. They assumed a transmission of stress by tension to the roots, which, because of their angulation, was more or less expended as ecrassurage (crushing), at the points where the roots angle around the transverse processes. They held that this action was enhanced on the lower roots by elevation of the arm, and especially on the first dorsal root as it wound around the neck of the first rib. They believed also that the force was transmitted far back even to the spinal marrow in most cases. In dislocation they believed that traction fell on the roots at the same time that it caused the dislocation, rather than that lateral stress was caused by the dislocating head. Most of the French and German profession accepted this explanation, and the idea that a supraclavicular, even radiculo-medullary lesion took place in most of these brachial plexus injuries became more and more engrafted in the minds of surgeons and neurologists. It remained for Delbert and Cauchoix to point out in a very wonderful paper in the Rev. de Chir., in 1910, that many of these supposedly supraclavicular lesions that had been diagnosed as radiculo-medullary, root or trunk lesions were in reality lesions of the axillary portion of the plexus; that is, terminal, and especially so when they accompanied dislocation of the shoulder. They reported three operated cases of their own, and the rest were interpretations of cases previously reported by others; but only a few were checked cases. They believed that the injuries were high up, i.e., root lesions, or low down and terminal, but not of the plexiform part. In this last conclusion we agree. They studied thirty-six cases, all dislocations; and they believed, as did Duval and Guillain, that the nerve lesions were due to the same stresses which caused the dislocation, rather than to lateral stress from the displaced head of the humerus. Their paper never has received the attention in America to which it is entitled. It is probably true today, when one speaks of a brachial plexus paralysis, that the majority of American surgeons visualize a lesion of the nerves of the plexus high up above the clavicle, either trunk or root or even radiculo-medullary, and give little thought to the possibility of infraclavicular peripheral injuries, especially to those of the axillary nerve. There has been, we believe, a misconception in the minds of the profession regarding the frequency of brachial plexus rupture. True rupture, with complete separation of the torn ends, is a rare lesion either above or below the clavicle, whether of root or terminal branch, as compared to cases of injury without rupture. The cases of proved rupture, although few, are extremely important, both as to their mode of production and as to the exact point of the plexus which has suffered the separation, because from a study of these proven cases, it is easier to understand where the acme of stress is most likely to fall, in those less severe and much more common cases where paralysis supervenes without actual loss of continuity in any part of the plexus. It has been said by various authors that Bowlby was able to collect nineteen cases of frankly supraclavicular rupture of the plexus, that Bristow added three more proven cases, and that Frazier and Skillern were able to collect records of twenty-one cases of actual supraclavicular rupture. Examination of all these original papers shows that Bowlby reported nineteen cases of brachial plexus paralysis, but only three of them were checked, namely, one case of his own, where operation was in the axilla and nothing was found; the case of Banks of actual proven rupture, which had already been reported; and the old case of Flaubert of actual avulsion, which has been reported so many times. Bristow reported one case of his own of actual rupture, and another case of actual rupture which had already been reported by Hartley. His third case was by no means proven; he attributed it to Senn because Senn had told him that he (Senn) had had a case. No details of this, whether proven or not, were included. Frazier and Skillern reported only one case of their own of actual rupture, and the celebrated case also reported by Mills; their others were cases from literature without detail, and again included the case of Flaubert. Frazier simply said that in the literature he had found records of twenty-one cases with proven rupture of one or more roots, without skeletal injury. It may, therefore, be understood upon how little foundation theories as to the causation of brachial plexus ruptures now rest.

DR. STEVENS' THEORY OV MECHANISM OF PRODUCTION OF BRACHIAL PLEXUS INJURIES

A cord to be broken in tension, whether by a blow on its side transmitting the stress to both ends, or by direct pull, must be held firmly at the ends or there will be no tension. The brachial plexus is so held by the fascia, as we have shown in our dissections and by the description already given. All cases of plexus injury of the type under discussion are due to tension, that is, traction. It makes no difference as to the basic mechanical stress whether a man slips carrying a heavy load; whether he is struck by a blow, depressing his shoulder; whether there is straight traction on the arm; whether his head is forced away from his shoulder; whether his face is rotated away from or toward the side undergoing stress; whether the arm is pulled in abduction and in external rotation; whether or not his arm is raised, lowered, supi-nated or pronated; if the stress is reflected on the cords of the plexus, it is tension. However, the different varieties of stress and the relative position of the ^rm and head at the time of stress, make tremendous differences in the kinds of lesion suffered, in the locality of the lesions and in prognosis. The nature, location, and seriousness of the injury, I believe, depend on whether or not the stress is received from above or is transmitted from below; whether or not the arm is above or below horizontal, and whether or not externally or internally rotated. These factors, together with velocity and magnitude of stress, determine, in most cases, the severity and the locality of the injury. Nevertheless, in all cases the stresses are of the same nature.

Consider the cords of the brachial plexus as a traction apparatus with its normal axis as a mechanical appliance on the 7C vertebra, with the arm at the horizontal, i.e., a single cord with five separate points of attachment firmly snubbed at the transverse processes. When tension is applied to this structure (the integrated cord), it falls on the offset roots. Any mechanical engineer will tell you that an arrangement of this kind will rarely transmit stress through five cords equally. If the force of the pull could fall exactly through the neutral axis at the exact center, and at an exact right angle to the base or plane of the structure to be lifted, the size of the cords being the same, it might be possible to lift a weight evenly, but he will also tell you that stress always tends to travel in straight lines, and depending on the position of the application of stress, the acme will usually fall to one side or the other of the neutral axis of such a structure.

A suspension apparatus is governed by much the same laws as a traction apparatus. In engineering a three-point suspension is more reliable than a suspension from a greater number of points. Perhaps this is the reason why three roots are injured in so many cases of brachial plexus paralysis, for either the two upper roots or the two lower ones may combine with the median root. A traction apparatus must have a neutral axis and a hne of resistance, and when the force of traction falls through this neutral center or axis, the traction is equally borne by all parts of the apparatus. Even a slight deviation from this neutral axis makes an offset pull to one side or the other, and in a structure of this kind, if the line of tension falls outside the neutral axis, as represented by the line of resistance, the entire force is transferred from that neutral axis; all tension is released on the cords on the other side, and a new neutral axis is instantly formed about the components which are now bearing the stress, to conform to the new line of resistance. All other components are out of the structure; they are lax and their influence is nil. A pulley inserted as part of a traction apparatus is not placed so as to change the degree of pull on the structure to be lifted or moved. It is placed there in order to change the direction of the application of the force in order to make it more convenient or effectual, i.e., to keep the neutral axis in the desired direction. The pulley must be so placed that a line from the pulley to the center of the structure to be raised or moved, falls through the neutral axis and the line of resistance. If this is not so, the force applied falls to one or the other side of the axis and the entire force of the pull may, therefore, fall to one side. If you elevate the pulley, the lines of tension and resistance will come below the neutral axis as it existed at first, and the acme of stress will be below. If you lower the pulley, the acme of stress will be above. Since the scapula is movable and the integrated cord passes under the arch formed by the coracoid and the pectoralis minor, a condition similar to a movable pulley exists in the shoulder. There is no real pulley, but the cords of the plexus are held in this arch and, as the arm is raised and the clavicle and the coracoid rise, the latter acts much like a pulley, for it changes the direction of any force applied distal to it. In raising the arm, when the coracoid rises above the horizontal, the acme of stress would come on the lower roots. (Fig. 64c.) Lowering the arm lowers the coracoid, and the acme of stress is on the upper roots. At a horizontal, the neutral axis is near the 7C vertebra, and all roots might be stressed evenly, because the neutral axis, the line of tension and the line of resistance correspond. As a matter of fact, when the continuity of the bones is intact, the range of the coracoid is such that like a self-adjusting pulley it maintains the direction of pull very nearly in the neutral axis, as the arm moves.

As an exaggerated example of how the coracoid can change the direction of stress, suppose a man caught in a machine in such a manner that, while the arm is abducted in the scarecrow position, two parts of the machine travel in opposite directions, one pressing the elbow upward and the other part pressing the shoulder downward. The coracoid will be depressed, hooked over the taut plexus, and the force transmitted to the upper roots. The stress would be downward on the upper roots, even if the pressure above the shoulder remained fixed and only the elbow continued to be forced upward. With the arm at the side and pulled downward, the pulley is not the coracoid, but the place where the plexus comes over the first rib anteriorly. This slight change of direction of the force would relieve, to some extent, the strain on the upper roots in a downward pull* Combined with this would be help from the clavo-pectoral fascia. A breaking strain expended on the brachial plexus from above, as from a blow on the shoulder, or the stress suffered when one slips while carrying a weight, usually should cause a lesion of the 5C root. Five cords divided will not stand the strain as well as when they are combined in one. The apparatus will break at the weakest point, i.e., at one of the roots between the point where it is snubbed on the transverse process and the junction of that root with others. It breaks there, or it does not break at all in supraclavicular stresses. If it does not break, nevertheless the acme of stress is at the same point. In this case of a blow from above on the neck or shoulder, or a slip while carrying a weight on the shoulder, "the integrated cord is stretched, and the stress in both <sases is transmitted to the points of firm attachment of the cord at both ends. Here is impact, and impact is infinitely greater than static load. A man carrying one hundred pounds on his shoulder slips or makes a false step, and falls even four to six inches. Instantly his static load, which he could carry with ease, becomes an impact trauma. The mechanical stress suffered here is exactly the same as it would be if he were struck on the shoulder by a hundred pound hammer falling through four to six inches and depressing the clavicle and the coracoid. The brachial plexus is instantly stretched between its two firm points of attachment, which are, as we have shown, at the transverse processes above, and at the clavo-pectoral fascial snubbing below, in the upper axilla. Again, the stress in tension is exactly the same as if the cords themselves received a side blow. Impact, as the mechanical engineer will tell you, so increases a stress over static load as to be almost unbelievable. (See Merriman, Mechanics of Motion.) When surgeons contend that rupture of the roots would be impossible in the ordinary trauma, or in that due to dislocation, they fail to take into account the manner in which mechanical stresses may be magnified at the point of final application when they fall on structures of limited area. The reason that the roots do not always break, is that the stress does not fall on these limited structures alone in many cases. It is disseminated. In both cases cited above, the stress is transmitted to both points of attachment. One half is referred back to the roots, the other half falls on the place below, where the cords are held firmly by the clavo-pectoral fascia. So that sometimes, even in this type, we have an accompanying rupture or injury to the artery, because the fascial snubbing surrounding the cords gives way and the vessel is torn. Injury to the artery probably happens less often in case of blows from above than when the stress comes through the arm from below as pure tension, because in the latter case, the entire force of the pull is not divided. It falls first on the clavo-pectoral fascial snubbing, and is then transmitted back to the pulley at the first rib and the roots, but not until the fascial snubbing in the axilla has been injured. It is quite possible that we might have traumatic aneurism without rupture of the roots, because the stress might be disseminated after the fascia and the vessels have been injured, and then be too weak to break the roots.

In a very careful review of the literature the writer has been unable to find a single case of rupture in the plexiform part of the plexus, i.e., in that part which has been alluded to as the neurovascular cord, although the artery itself may be torn. Even proved injuries of this portion of the plexus from bullets or cuts are rare, but we are speaking now of rupture from trauma without a penetrating wound. In reading the accounts of operations, on the other hand, one usually finds such statements as "the plexus seemed a mass of scar tissue," "the cords were welded together in an inflammatory mass," "on account of the scar tissue, nothing could be made out as to the exact location of the injury," etc., etc. Yet many such cases have recovered in whole or in part, indicating that no real rupture had occurred in the nerve fibers, and that the gross appearances were due to ecchymoses, exudate, or scar amongst the fascial envelopes and fibrous septa in the nerve trunks. In this region, as in others, intensive anatomic study of the mechanics of the structures reveals marvellous examples of architectural and mechanical designs, e.g., the integrated cord when dissected is "plexiform," after the fashion of a complex design of parallelograms of forces. This is an admirable arrangement to disseminate stresses, for if the cord were pulled at both ends, before rupturing the longitudinal strands, the force must break the little lateral bands of tissue which we cut when we dissect the plexus. It is probable that many little local hemorrhages, about, amongst, between, and within the individual trunks, cause the appearances so frequently described. Subsequent exudate and scar tissue complicate the picture, and choke the nerve fibers, usually temporarily only. I believe the explanation of these injuries to the plexus is this simple mechanical one; in any case it has nothing whatever to do with the idea of ecrassurage against either rib or transverse process. The stress will not be "radiculo-medullaire," except in very rare cases of sudden, violent injury, because to be so, the mechanical apparatus would have to break back of the snubbing, and that is contrary to the laws of mechanics. The snubbing is stronger than any individual root, but it is not stronger than the entire integrated cord. If the stress comes from below through tension on the arm, there must be some separation of the bony framework at the shoulder, before stress can be transmitted to the soft tissues. For instance, the clavicle might be broken, or the humeral head might be pulled away from the glenoid cavity, or when dislocated, pried against the plexus. In postoperative paralysis in breast cases, the muscles are relaxed and the plexus assumes the stress. In cases of fracture of the clavicle when paralysis is a complication, we should think first of a root lesion, and only secondly of injury to the cords from actual contact with fragments.

Estimate of Comparative Strain on Integrated Cord and Root.

Some idea of the strain to which a root is subjected in one of these injuries, in comparison to that sustained by the integrated cord, may be obtained by reference to Fig. 64. If the pull remains the same the stress per square inch becomes rapidly greater the smaller the cord. Engineers use the formula

Pull / Pi R (squared) = stress per squar inch

By using this formula and assuming the integrated cord comprising the artery, fascia and nerve trunks to be one-half inch in diameter, we find that a pull of one hundred pounds would exert over five hundred pounds stress per square inch on the cord as a whole. If all the other components were removed and this stress fell on one nerve root one-tenth of an inch in diameter, this root would have to receive over 12,000 pounds of stress per square inch of its cross section, which, of course, would rupture it. If the stress fell on two roots, it would be divided. In order to determine the comparative stress on these two roots, we use the parallelogram of force of the mechanical engineer, Fig. B. Therefore, if it is a breaking stress, the roots never break together; one gives, and then the entire stress falls on the next, etc., until dissemination of force stops the process. With stresses as above, why do these cords not always break? Simply because these stresses do not often fall on the cords in this way. Other structures, bones, ligaments, muscles, tendons, fascia, etc., receive some stress, but the bulk of the stress may thus fall on one root or two roots, and sometimes does.

Considerations such as these make us wonder at the remarkable arrangements which nature has supplied to prevent stress falling on any one root of the plexus. When a mother drags her child across the street by one hand, there are many anatomic structures which protect the plexus. The bones and ligaments form a chain which takes most of the stress. Should there be a loss of continuity in this chain, the coracoid process acts as a pulley to distribute the stress in the neutral axis of all the roots. Even when the stress arrives at the roots, the two lower ones join together and the two upper roots join together before they unite with the middle root; thus, the upper cord and the lower cord each form a two-point suspension, and the middle root is somewhat protected. When the two points formed by the upper and lower pairs unite with the 7th or middle root, we have a three-point suspension. These three-point suspensions then form the plexus, which is so arranged that each main cord and nerve is attached to its neighbors by fascial binding, uniting them into a trunk capable of sustaining much stress. Even then, when they pull apart, the fascial bindings between the cords must be torn before any individual cord may be torn. However, there are occasions when, owing to rupture in the continuity of the supporting structures, all stress will be thrown on either the uppermost or lowest root. Even then there remains the snubbing of the fibrous envelope of the nerves to the transverse processes to protect the spinal portions of the roots. Fainting, anaesthesia, or other forms of unconsciousness which relax the muscles remove a very important factor in protection.

The Theory Applied to Ruptures of the Terminal Branches.

Let us now consider the stresses which often result in terminal branch lesions. Imagine an arm in abduction and external rotation subjected to still greater backward stress,* This is the position of dislocation, and lateral pressure on the taut plexus takes the place of pure tension on the whole arm structure. By this position of abduction and external rotation you have separated the terminal branches of your cord as much as possible, and if lateral pressure is applied, your apparatus is no longer a single cord with five lesser points of attachment where it is joined to the transverse processes; it now has six or eight separate smaller cords below, any one of which may have to take the major stress. In other words, you have added other factors to your mechanics. Offset may now fall, not alone on the roots, but at the other end of your apparatus. If it falls on any terminal branch that is smaller or weaker than any part of the integrated cord or roots above, it will break that terminal branch below its point of insertion into the integrated plexiform part. If, however, the main stress should fall through the radial nerve, which is greater in strength and size than most of the roots, it would probably be transmitted back to the roots as the weaker points. I have found in many dissections that the posterior or intermediate cord is as high in the neck as the lateral fasciculus. Force applied to the radial would be transmitted through the integrated cord back to the 5C and 6C to a greater degree than to the 7C, from which its motor portion arises originally. The weaker points are below and at the roots. This is why there have been ruptures reported of the musculocutaneous nerve, of the median heads and of the ulnar, at their origins in the axilla. We have many reports of injury involving each of these nerves in the axilla (forty-four operated cases), but there is no proven case of absolute rupture of the radial alone. Why? Because it is stronger than the clavo-pectoral fascia which binds it, and while it is weaker than the plexiform part of the plexus, and would break before that, it is stronger than the roots themselves on which individually, because of offset, the stress may also fall. When we speak of a root, or a cord, or a nerve, as stronger or weaker, we mean that cord or nerve bound with connective tissue. No root or cord breaks until after its sheath has given way. No nerve can be stressed -until its surrounding fascia gives. The fact that fifteen cases of subclavian or axillary aneurism have been reported in conjunction with 135 injuries to the plexus, suggests that the clavo-pectoral fascia surrounding the integrated cord is usually stressed and often actually torn. In most of these cases there was only one incision, either a supra- or an infra-clavicular one, so that the conjunction of high and low lesions is probably greater than indicated in these figures. If in each case both incisions had been made, a still greater coincidence of fascial and arterial injury might have been shown.

• In this position the terminal branches tend to separate from one another, and in the opposite position (internal rotation with the arm at the side) to continue their course almost side by side.—E. A. C.

Application of These Principles to Axillary (Circumflex) Nerve Injuries.

In dislocation of the head of the humerus, we have the cause of the greater number of axillary nerve paralyses. We have at times in addition injury to each of the other nerves in the axilla, and through the radial, for the reasons stated above, we may also have as a result of the stress of the rotating head, injury transmitted back to the roots. The axillary nerve is most often injured because in dislocation it is tensed across the head in the erect phase when the arm is elevated and externally rotated. The axillary is the one nerve which at this point has already separated, and is an individual cord. It is closely applied to the anterior surface of the subscapularis muscle, and is carried with the head when it turns out of the lower part of the glenoid. Since the subscapularis tendon is anterior, and firmly attached to the lesser tuberosity, this nerve must be more or less stretched in every case of anterior dislocation. The axillary is particularly vulnerable because it is firmly attached above to the main cord, and is snubbed below, where it passes back to wind around the neck of the humerus. Therefore, it is a short length of nerve subjected to a side stress, which is again nothing more than tension transmitted to both ends. The wonder is that it is not always paralyzed in cases of dislocation. Confined solely to the axillary nerve would the stress be transmitted back to the roots? It would not, for the axillary is weaker than the plexiform part of the plexus, and it is weaker than any root. The axillary is stressed most often alone, not enough to rupture it, but enough to paralyze the deltoid. If it breaks, it will break between the place where it leaves the intermediate fasciculus and the place where it passes between the subscapularis and the teres major muscles, and nowhere else, because this is the weakest point. This accounts for the fact that simple deltoid paralysis is far the most common nerve lesion complicating dislocation.*

• In such cases the patient might still be able to elevate the arm with the supraspinatus.—E. A. C.

In our series from the literature, there were only three cases of actual rupture of the axillary nerve, aside from many cases of injury where the exact pathology was not determined by the operation. In addition, Weir-Mitchell reported nine cases of ruptured axillary, years ago, but they were not checked by operation.

The cases of musculocutaneous paralysis (aside from rupture), which have been reported, may be, and sometimes are, due to the mechanics already described. We believe they may be caused also by the pressure of a dislocated head which has been left out for a long time, although the nerve may be stressed by the rotating head in dislocation, especially in the first phase, exactly as all the others may be stressed. The idea that these nerves slide up and slip over the head of the rotating bone with any degree of facility is absurd. The only way they slip over the head is because the head of the bone is rotated under them, in spite of tension, and in spite of their being plastered in place by the tremendous contractions of surrounding muscles which are resisting injury. This nerve is also snubbed at its junction with the integrated cord above, and it is snubbed where it passes through the layers of fascia both internally to the biceps and beneath it, and where it usually passes through the coraco-brachialis muscle. Like the axillary it is a short cord which is being tensed by a side stress. It may be injured in this way or even ruptured, but we believe that many times as the dislocated arm falls back to the side, the musculocutaneous nerve, in this new position, is unable to free itself, and remains tensed over the dislocated head, until reduction frees it.

A case of subcoracoid dislocation of the right shoulder, caused by a fall, was observed by the writer in 1929. It was three hours before the dislocation was reduced, during which time the patient swung his arm, because he said that although it hurt him to do so, it felt better when he did it. He had a sensation of pins and needles during this time in arm and forearm. Reduction was not followed by abduction treatment, but the arm was supported by a swathe and sling. When he returned four days later, paralysis of the biceps, brachialis and coraco-brachialis was found. Two weeks later the reaction of degeneration was present in these muscles. No others were paralyzed. Epicritic loss was greater than protopathic in the domain of the musculocutaneous nerve. The absence of brachioradialis and supinator brevis paralysis, the absence of supraspinatus, infraspinatus, deltoid or teres minor paralysis, the presence of coraco-brachialis paralysis and the fact that paralysis did not occur at once, make it probable that this was an injury of the nerve in its axillary portion. It is best explained by compression from infiltrate which came as the result of long-continued pressure. First voluntary motion appeared at the end of eight weeks, and recovery was very rapid thereafter.

Although they are very common, very few cases of isolated deltoid paralyses, or combined paralyses of the deltoid, biceps and brachialis, appear in the literature, because many of them are mild and recover in a few months. Nevertheless, I believe that axillary injuries of the brachial plexus are much more numerous than supraclavicular injuries, and also, in the main, less serious. In this instance, ordinary practical experience is of more value than recorded literature. Lesions of the plexus caused by, or at the time of, dislocation usually soon get well; cases in which recovery is prolonged or eventually imperfect are those in which the shoulder as a mass has been suddenly depressed. Thus I believe that with a knowledge of the actual trauma which was sustained, it is possible to arrive at a better conclusion as to prognosis than it is from the clinical symptoms alone.

Avulsion of the nerve roots from the spinal cord ("radiculo-medul-laire") must be very rare. There are only four cases of proven avulsion which I have been able to find in the literature and I shall speak of these later. In these four cases at least there can be no doubt of the lesion. How can we explain the mechanics of the avulsion cases, if, as we have said, the anatomical arrangements predispose toward more peripheral injuries? The simplest explanation is this: The mechanical engineer will tell you that in the face of tremendous stress, where velocity of stress is great, all rules governing stresses and strains fail. In the face of a tremendous stress with velocity, the snubbing of the roots at the transverse processes would give way and the stress be then applied to the nerve fibers within the vertebral canal. It is significant in these four cases that in not one was the 5C broken. It is evident in all that the stress fell through the 7C and the 8C, which are certainly the largest of the roots. The cause in each case was a direct pull in nearly the neutral axis. It was pure tension by pulling upon the arm in four cases, and in three of the four cases, at least, it was a tremendous stress, and probably so in the other. In all cases the arm was at or above a horizontal. In the case of Frazier, the arm was raised and the element of pure traction was evoked through the lower roots. In no other way can the escape of the 5C root be understood. In our opinion, there would always be in any real case of avulsion, mild or severe, cord symptoms which would be diagnostic. A modified syndrome of Brown-Sequard should be present below the spinal level of the lesion.

Detail of Nerve Paths in the Plexus.

Now that we have seen how traction may be distributed and, through the integrated cord, affect one root or another according to the direction of the pull, it is time to consider the detail of the nerve paths through each root and cord to individual muscles. Such knowledge as we have of the distribution to the various muscles of the arm of the nerves which form the plexus, has been obtained with much labor in the past by those who approached the problem in three different ways, namely:

   (1) Dissections on the cadaver to determine the exact anatomic pathways of nerve fibers from the spinal cord through the intricacies of the plexus to each muscle.
   (2) Study of the cases recorded in the literature where definite paralyses of the different muscles have been recorded, and later were shown, by operation or autopsy, to be due to definite lesions of the plexus.
   (3) Experimental studies in animals where lesions were made by operation and the resulting paralysis in the muscles carefully recorded, or where electric stimulation at different points in the plexus caused contraction of individual muscles.

Much of our knowledge has come from studies made by Paget, Jonathan Hutchinson, Dej erine-Klumpke, Herringham, Seeligmul-ler, Erb, Duchenne, Bardenheuer, et al., in the era prior to the general adoption of antiseptic technique in surgery, a discovery which so broadened the interests of medicine that other fields more promising than plexus injuries were opened up. However, there is always some one interested in every problem, and to this one Harris, Sherren, Sherrington, Claude, Marie, Thomas, Henri Meige, et al., have made notable contributions in more recent years. It has interested me to make a thorough review of the literature, to collect and study as a group all of the reported cases I could find, and to compare the various diagrams of the plexus presented by different authors, especially those of Kocher, Harris, Kerr, and the many anatomists. While the diagrams present a notable unanimity in general, there is much disagreement in detail, owing to the fact that there actually exists a considerable degree of anatomic variation. To satisfy my own mind I have made ninety-two dissections of the plexus and have recorded my findings. Fig. 65 shows what I believe to be as accurate a diagram as can be made of a structure which is subject to so many minor variations and which, as those of us who have made dissections best know, may be so easily altered or distorted by our own manipulations.

FIGURE 65

This chart was made personally by Dr. Stevens, as well as the notes which accompany it. I have not attempted to correct, or in any way change, the chart, although the terminology is not the one in use by anatomists at the present date. Dr. Stevens was accustomed to the old terminology, but in some instances used the more up-to-date one. I believe that any earnest student of this subject could change these terms as accurately as I might. If I attempted corrections I might make a mistake, for I am not as learned in the subject as was Dr. Stevens, although there are some obvious errors in minor details. As the notes were unfinished I have been obliged to correct them to some extent, although I have not checked them with the writings of the authorities quoted.—E. A. C. In this scheme only dominant roots, or branches which have a large part in the innervation of the individual muscles, are given consideration.

NOTE NO. 1.—The Coraco-brachialis, according to Piersol, is supplied from the 7C root; the Manual of Neuro Surgery, issued by the War Department, 1919, states from the 5C root. Harris in his "postfixed plexus" gives an 8C branch, but there is no record in the literature of a lower root injury accompanied by paralysis of the coraco-brachialis. I have never seen such a connection in the cadaver. Braus gives 6C and 7C. Spalteholz gives 6C and 7C, with 7C dominant. I believe 6C and 6C in many cases (and when so, the 5C and 6C are dominant), but more often it is from the 7C, in which case the 7C.becomes the dominant root. Many times it comes from the 5C, but probably usually is from the 7C via the pathway to the lateral fasciculus. At times there is a branch from the intermediate fasciculus to the coraco-brachialis by another pathway, as given by Harris.

NOTE NO. 2.—The Manual of the War Department gives the Adductor Pollicis as from the 6C and 7C root, which is, of course, quite impossible, except via a lateral ulnar head. Other anatomists give the 8C and ID, which is probably right, and to this we adhere. Stewart gives 8C and ID; Herringham, 8C; Harris, ID; Braus and Spalteholz, 8C and ID. The dominant root is the ID. See Recaldoni and Pfeiffer and Ransom cases.

NOTE NO. 3.—Pronator Quadratus. War Department gives a 6C in addition to the 8C and ID, and Piersol gives the 7C in place of the 6C of the War Department. Herringham gives 7C, 8C and ID. Braus, 7C-8C and ID. I am convinced that it does not come from the 7C. 8C is dominant and the influence of ID is little, if any; therefore it is not given on the chart.

NOTE NO. 4.—The 1st and 2nd Lumbricales are given by Piersol as 6C and 7C. War Department only the 7C; Herringham, 8C; Stewart, 8C and ID; Harris, ID; Spalteholz, 7C-8C-1D; Braus, 8C and ID. Our scheme is the 8C and the ID, for all the lumbricales, and ID is dominant for the 3rd and 4th lumbricales.

NOTE NO. 5.—The War Department gives the Plexor Carpi Ulnaris as having a 7C branch. If so, it would have to go via the decussation to the lateral fasciculus, and could not possibly reach the ulnar, save by the lateral ulnar head. We have, therefore, in our scheme given the 8C and ID (perhaps). Influence of ID must be little, if any, so it is not included on the chart. The lateral ulnar head in our opinion is only sensory from the 7C. Braus gives 7C-8C and ID; Stewart, 7C; Harris, 8C and ID; Herringham, 7C-8C-1D; Spalteholz, 8C and ID.

NOTE NO. 6.—The Flexor Longus Pollicis is given by the War Department as GC and 7C; as 8C and ID by Piersol; Herringham, 7C-8C and ID; Harris, 8C and ID; Stewart, 7C; Spalteholz and Braus, both 6C-7C-8C. We have adhered here to the 8C. No cases of injury to the lateral fasciculus or upper roots have ever been reported as showing injury to the flexor longus pollicis; and where the ID has been involved, the flexor longus pollicis has usually not been included. Notes continued on back of chart.

NOTE NO. 7.—The Interossel are given by the War Department as having A 7C branch, which we believe impossible, save in rare cases where the 7C gives a branch to the medial fasciculus. Stewart gives 8C and ID; Herringham, 8C; Harris, ID; Spalteholz, 7C-8C-1D; Braus, 8C and ID. I believe 8C and ID is more likely to be correct, ID being the dominant root.

NOTE NO. 8.—The War Department gives the (Flexor Brevis Minimi Digiti, O. T.) Flexor Brevis Digiti Quinti, B. N. A., and the Opponens Digiti Quinti (Op-ponens Minimi Digiti, O. T.) as having a 7C branch, which it seems to me is unlikely. Spalteholz and Braus both give 7C-8C-1D. We favor 8C and ID, and ID is dominant

NOTE NO. 9.—Flexor Brevis Pollicis. The inner head is innervated by the ulnar. In our scheme this head is considered as one of the interossei; therefore inner--vated as the interossei. The outer head is possibly 8C, and, if so, the nerve-path must go via the inner head of the median. Spalteholz, 7C-8C-1D; Braus, 6C and 7C; I favor 8C and ID.

NOTE NO. 10.—Palmaris Longus. The War Department gives 7C, 8C and ID; Piersol gives 6C. Probably 8C takes a straight course to the median, and this is the scheme to which we have adhered. The flexor carpi radialis is given in the Tinel Syndrome as of the outer median head. The flexor carpi radialis and the palmaris longus are innervated probably from the same roots, although it is certainly not the 5C and 6C. They may at times be supplied from the 7C, via the outer median head, but Marie and Meige in electrical stimulation of the outer median head in sixty-four cases caused contraction only of the pronator radii teres (O. T.); of the inner head, they caused contraction of the flexor carpi radialis, the palmaris longus and the other flexors of the wrist and fingers. From our analysis of cases we believe the dominant root for these two muscles to be the 8C, and we so give it. If at times it is the 7C, then in those few cases the path might go via the outer median head, but we believe that this must be rare. The flexors and extensors of the wrist are given by Stewart as 6C, and of the fingers as 7C, which we believe incorrect. Harris gives flexor carpi radialis and palmaris longus as the 8C, as we do. Herringham, 7C-8C-1D, which is unlikely. Spalteholz, the 7C-8C-1D for palmaris longus, and the 6C-7C-8C for the flexor carpi radialis. Braus, the 7C and 8C and ID for the palmaris longus, and the 6C-7C-8C for flexor carpi radialis. Our scheme gives the dominant root for the palmaris longus, the flexor carpi radialis, the flexor sublimis and the flexor profundus as the 8C. If there is a 7C branch, at times, it is not dominant usually, and neither is the ID, which, nevertheless, probably gives a small connection. Therefore, only the 8C is given in our chart.

NOTE NO. 11.—Latissimus Dorsi. The War Department gives 6C and 8C. Piersol is probably more correct, 7C and 8C roots. The only difference would be the presence of latissimus dorsi fibers in the cephalic division. Otherwise the course is the same, but the main nerve is the 7C. In many of our ninety-two dissections the 5C and 6C could have sent no branch to the thoraco-dorsal nerve. Purves Stewart gives 7C, which we know is not so, except in part; Harris as 8C-1D in postfixed, but he gives a 6C in prefixed; Herringham, 6C and 7C; Spalteholz and Braus both 6C-7C-8C. Our scheme gives 7C and 8C.

NOTE NO. 12.—The Teres Major may be innervated by way of the axillary, and this is sometimes the case. The upper subscapular, from the 5C and 6C and posterior division of the 7C, usually goes to the upper part of the subscapularis. The middle subscapular goes to the subscapularis also. The thoraco-dorsalis goes to the latissimus dorsi, and the axillary subscapular supplies the teres major'. Spalteholz gives 5C-6C-7C for subscapularis, and 6C and 7C for the teres major. Our scheme is the same for the subscapularis, but the 7C and 8C for the teres major.

NOTE NO. 13.—The radial extensors of the wrist, it must not be forgotten, are supplied sometimes by the 5C (see Harris and Low, Fairbanks and Sherren), but more often probably from SC-6C and 7C. In an analysis of the twenty-four cases of cut 5C, or 5C and 6C roots (not tension), the extensor carpi radialis longus and brevis were involved, or reported involved, only seven times, so they are not always from the 5C and 6C, therefore probably 7C; Herringham, 6C and 7C) Harris, 5C and 6C; Stewart, 6C; Spalteholz, 6C-7C-8C; and Braus, 6C and 7C. Our scheme is the 5C and 6C at times, and these are dominant. When from the 7C, as they often are, the latter is the dominant root.

NOTE NO. 14.—While many muscles possibly have a small branch supply from the 6C, it is generally admitted that there is little extension of the paralysis produced by involvement of the 6C when the 5C is already involved. Direct stimulation of 6C causes contraction of the clavicular portion of pectoralis major, posterior part of the deltoid and some of the forearm muscles, probably pronator radii teres. Stewart gives the pectoralis major as 6C-7C; Harris as 5C-6C; which it certainly is at times. Spalteholz gives it as the 5C-6C-7C-8C; and the pectoralis minor as the 7C and the 8C. Our scheme gives the pectoralis major as the 6C-7C and 8C, and the pectoralis minor as the 7C and the 8C.

NOTE No. 15.—All the flexors and abductors of the thumb, and all the extensors of the thumb, are given by various authors as having a 6C branch. Both Piersol and the War Department give the abductor pollicis as 6C-7C, as does Braus. Spalteholz gives the thenar muscles as the 6C-7C and 8C, and even ID for the opponens pollicis. Tinel specifically says paralysis of the abductor pollicis, the opponens pollicis and the flexor brevis pollicis is evidence of injury of the inner median head and therefore from the caudal roots. Very rarely have any injuries to the lateral fasciculus or cephalic roots been accompanied by disturbance of thumb action, and for that reason we have not given a 6C origin to any of these nerves. The flexors and abductors of the thumb are innervated by the 8C and ID, through the inner head of the median, and the extensors of the thumb by the 7C and 8C, dominantly at any rate. Andr6 Thomas and Leverty are the only authors mentioning an involvement of the thenar muscles in an upper root lesion. Braus gives the opponens pollicis and flexor brevis pollicis as 6C and 7C, and the adductor pollicis as 8C and ID. Our scheme is 8C and ID for flexors, abductors and adductors of thumb; for extensors of the thumb 7C and 8C, with a small connection from ID at times, which is not dominant, and therefore not included in our chart. The abductor pollicis longus (B. N. A.), it must be remembered, is the extensor ossis metacarpi pollicis of the old terminology as it is called in this chart. Its innervation is via the radial from the 7C and 8C, and it is considered here as an extensor. Nevertheless, in a pure median paralysis through its action, the thumb may still be abducted.

NOTE No. 16.—The Pronator Radii Teres is often supplied from the 5C or the 5C and 6C. Many times probably from the 7C, when not from the 5C or 6C. Sherren says that in his Erb-Duchenne type he never saw the pronator radii teres involved. In two of Harris' cases it was involved, and in two of Andr6 Thomas'. These were skilled observers. In our series it was involved with a 5C, a 5C and 6C, and a 6C cut, so it certainly comes from the upper roots, at times, but that it does not always do so, is evidenced by our list. In Apert's case with a 7C-8C-1D involved, all the epitrochlear muscles were a mass of fibrous tissue, While the epicondyiar were in good condition. So probably the pronator radii teres is at times from the 7C. Braus gives 6C and 7C.

Animal experimentation has been of some value, especially on monkeys, but the differences from the human have prevented very much advance from this direction. I do not feel that I have myself derived enough benefit from articles on this phase of the subject to review them here. The anatomic method giving the best results has probably been the examination of the macerated fetal structure and the tracing of individual nerve fibers through the intricacies of the plexus. This was the method used by Herringham. The anatomic charts which we have incorporated here have many points of difference from those which have been published by various authors and anatomists. The root scheme which we have found helpful, differs quite markedly from most of those published, including those of Tinel and Benisty, Henri Meige and Purves-Stewart among the neurologists, and of Piersol, Gray, Spalteholz and other anatomists. Some of the reasons for these differences will be found in the notes attached to the charts, or in the text. Figure 65 will give the reader our idea of the root innervation and of the ordinary arrangements of the plexus better than a detailed description, bearing in mind always that this representation is only schematic, and does not resemble in the least what the surgeon sees in dissection as he comes down upon it from the supra- or infra-clavicular incision.

Prefixed and Postfixed Plexuses.

There have been several classifications of the varieties of the brachial plexus. Harris's division into prefixed and postfixed types has been very generally accepted and used by anatomists and neurologists. Briefly, the prefixed type has a 4C connection to the 5C and no 2D to the ID, while the postfixed plexus has a 2D connection to the ID and no 4C connection to the 5C, therefore making them essentially different types, which do not overlap. This would be of little practical importance, if true, since even its authors do not claim that it raises or lowers the individual root constituents of the plexus one whole vertebral segment. There are other minor individual nerve difference;.! in this classification, but these are the essentials. I doubt the importance of this classification and give my reasons as follows. In our 1930 series of sixteen plexuses there were four with small connections from the 4C to the 5C root, as usually described. In three of these prefixed plexuses there was also a 2D connection making them, at the same time, postfixed plexuses; but both of these connections were small. This was not all. By careful dissection I found in eight plexuses another small connection, apparently from the 4C, which ran down through the foramen transversarium, beside the vertebral artery, to join the 5C root outside the transverse process. I was unable to find any record of such a nerve in the literature. I traced it into the spinal canal at the 4C. In no case did it come from the sympathetic trunk, so far as I could determine, and I cannot believe it to be a fiber which comes out with the 5C root beyond the transverse process and then turns back to become part of the sympathetic nerve, which constitutes the so-called vertebral nerve of the sympathetic, described years ago by Vulpius and Francke. I regard it as another small connection from the 4C to the 5C. In six of the eight, it was present with a 2D to the ID. In the two others, the 2D to the ID connection, it was probable, although it could not be determined absolutely, as against a connection with the sympathetic alone, since the relation was too close to permit of a decision. The 2D connection to the ID was sometimes separate, but usually connected with the sympathetic branches to the first thoracic ganglion. In several it was impossible to tell whether or not it was a 2D connection to the ID, or a branch to the ganglion. The thorax always has to be opened in order to be sure of a connection from the 2D to the ID. In the sixteen dissections a 2D connection was present in ten cases, and impossible to determine in four others.

I believe, if carefully looked for, there is in the maj ority of cases a small 4C connection to the 5C, and also a small 2D connection to the ID. This is of importance, especially since the operation of stellectomy has come so prominently to the front. Injury during removal of the stellate ganglion to this 2D connection to the ID, when it exists, may be accompanied by some minor paralytic involvement in the ulnar distribution in the intrinsic muscles of the hand.

I submit that with these findings, the idea of a prefixed and postfixed plexus must be rejected, because if a plexus is prefixed it certainly cannot be postfixed, even as a distinctive academic type. If we accept the presence of this vertebral nerve filament which we have described, we have ten plexuses out of sixteen dissections with a 4C connection of some kind to the 5C, and also ten with a 2D to the ID and eight with both. With this proportion, any classification based on the presence or absence of a 4C connection, or the presence or absence of a 2D connection to the ID, would seem to us artificial, unnecessary and misleading. The posterior connections to form the intermediate fasciculus usually come from the roots individually, but the amount of fascia left unremoved may account for different opinions. In the sixteen cases of the 1930 series freed from enough fascia, we found that they not only come from the roots behind their junction in the case of the 5C and 6C, but the same arrangement holds good for the 8C and the ID. However, in sixteen plexuses the ID failed to send a connection to the intermediate fasciculus six times. It was nearly always small, even when present. Cunningham originally denied any such connection, as did Testut and Quenu. Herringham states that it is seldom present, yet Harris found such a connection in eighty-two per cent of his cases, and Kerr found it in 169 of his 175 plexuses.

Classification of Anatomic Variations in the Brachial Plexus.

Obviously no such single diagram as Figure 65 can be utilized for every instance of brachial plexus injury, for almost every individual plexus would differ in some detail from any standard we might set. However, we may attempt to classify the most common and important variations in the hope that in time we may be able to understand exceptional cases of paralysis. These variations have been classified especially elaborately by Kerr, and for the sake of brevity, I present his study of the plexus here in a tabulated form, which I have myself arranged from his written descriptions.

This chart gives as comprehensive an idea of these different types as I can find, but we must assume, as he does, that the types do not overlap, although this is evidently not strictly true. His divisions are perhaps useful for anatomic study, but must be greatly simplified to be of use to the practical surgeon.

The normal type (Types B, E, and F of the chart belonging to Classes I-III) includes 93.7 per cent of all his dissections, and Classes I and II include 91.58 per cent, or all but thirteen cases in which there was a branch to the cervical plexus. This last is an artificial distinction, and to me it seems more simple to classify the plexuses as I have below. Certainly the normal type should be Type A to facilitate remembering the different types. Whether or not there is a small ascending connection, consisting of one or two filaments at most, to the cervical plexus above, on which Kerr bases his Class III, is a matter of no particular importance, even anatomically. In our ninety-two plexuses it was present in only thirteen per cent. In our 1930 series of sixteen plexuses, very carefully done, it was present in two cases, one a fair-sized nerve. Kerr found it in only seven per cent of his series. In ninety-two dissections, twenty-two of them superficial observations, but nevertheless including the roots, we have never seen a 4C connection to the 5C which was of the size described by Harris, and most of the connections were small indeed. I have never seen a brachial plexus with a 4C connection as large as that represented by Kuntz in his wonderful work on the Autonomic Nervous System, but that is simply a representation. I have seldom seen this connection of a size which would permit of a suture, as has been reported in the literature. In such cases we should gravely question the reports. Especially with scar and infiltrate present, we do not believe that any surgeon could suture this connection in most of the cases as we have found them. Kerr altogether ignores the presence or absence of a 2D connection to the ID and therefore makes no claims of special types depending on this connection. All his classes are subdivided into types which he designates by letters which do not correspond for the various classes and consequently tend to confuse. His types A and D, depending on a 7C connection to the medial fasciculus, ignore those plexuses which contain a lateral ulnar head, which we contend is probably normal or at least common, and which probably comes from the 7C root, although via the lateral fasciculus where this exists, and certainly goes to the inner median head and to the ulnar, which are the terminal branches of the medial fasciculus. If it goes to the terminal branches of the medial fasciculus, it would seem to be much the same thing as if it went to the medial fasciculus itself. In sixteen dissections we found it present eight times.

The individual variations of terminal branches which might be the cause of unusual paralyses are many. Of variations in the larger anastomoses in the upper portion, that of the musculocutaneous to the median is the most common. In seventy-five plexuses Kerr reports a connection from the musculocutaneous to the median eighteen times. In two of our last sixteen dissections there was a fused musculocutaneous and median as far as the elbow, and in six others a less extensive connection. In none of the entire ninety-two did we find a connection from the median to the musculocutaneous. Ignoring the smaller connections, in ninety-two plexuses we had eight large connections from the musculocutaneous to the median, where this connection was of such a size as to appear to be the main terminal branch of the musculocutaneous. In three, the junction of the musculocutaneous and the median was just above the elbow. In one, the median outer and inner heads were separate to the elbow, and at this point were joined by a branch of the musculocutaneous, all three being of equal size; therefore, three separate pathways for the median innervation. It may be seen that a peripheral nerve injury involving only one or two of these branches would be difficult to locate. In one dissection the entire musculocutaneous joined the median four and one-half inches below the clavicle. In only one case did we have a connection from the radial to the ulnar high up in the axilla, and in only one was there a connection from the median to the ulnar at the elbow. In the hand and forearm, the median and ulnar often sent small branches of anastomoses. A knowledge of these anomalies may be of use in any case of unusual muscle or sensory paralysis.

The types of brachial plexus, so far as stress is concerned, are three.

Type 1—where the suprascapular comes frankly from the 5C root or by two roots from the 5C and 6C before they join to form the lateral fasciculus. Obviously, a stress here could fall on the lateral fasciculus with a greater chance for the escape of supraspinatus and infraspinatus. There were fourteen such cases in our series of ninety-two, and by removing a little more fascia, it is found that the suprascapular goes back to the 5C root in nearly all cases, sometimes alone and sometimes with a filament or two from the 6C. In one dissection we found a suprascapular entirely separate from the lateral fasciculus, with three branches of origin—4C, 5C and 6C— these branches coming from far back on the roots. The origins from the 5C and 6C were at the exits from the gutters of bone at the transverse processes. In two cases of the entire series, but with a good deal of fascia removed, the suprascapular came directly from the 5C, and the 5C sent no ventral branch to the lateral fasciculus, but only a posterior which joined the posterior of the 6C and 7C. In this case the lateral fasciculus was clearly 6C and 7C, and the musculocutaneous nerve could contain no fibers from the 5C.

In the second type stress might well be transmitted so as to fall strongly or entirely on the 7C root. It is characterized by the fact that the 7C ventral branch passes directly to the outer median head. In fact, the ventral connection of the 7C is the outer median head, although it may receive a filament from the 5C or 6C. In the third type the 7C ventral division went to the medial fasciculus without any branch to the lateral fasciculus. In such cases any ventral 7C root fibers would have to go via the inner median head or the ulnar, and therefore it is clear that a wound of the medial fasciculus would cause more damage to the median innervation than would be usual, and might even involve the coraco-brachialis when this muscle is innervated by the 7C. There were only five cases of this type in our entire series of ninety-two. In no case of this type did we have a lateral ulnar head. So we assume, as stated above, that the lateral ulnar head is probably from the 7C. The last two of these three types were so small in number as to be negligible, except in an intensive anatomic study of the plexus. Therefore, in the majority of injuries, stress might not fall heavily on the suprascapular nerve.*

• In such cases the arm might be abducted by the action of the supraspinatus alone.—E. A. C.

Recorded Cases of Rupture or Division of Parts of Plexus.

The diagram of the plexus which I present as possibly an improvement on those of other authors has also been checked or modified from a study of 710 cases of brachial plexus injury from the literature. In arranging the root charts of muscle innervation, we have utilized the information obtained by an examination of the reports of paralyzed cases in which the exact pathology was known. There were some cases where, with cut or ruptured upper roots, it was possible to draw fairly accurate conclusions as to the innervation of muscles from those roots. A careful examination of the table, representing muscle paralysis, with cut or ruptured 5C, or 5C and 6C, or 6C roots alone, will, we believe, bear out our statement. Although these observations were mostly recorded by trained men, many were not neurologists trained in electrical reactions, and it is likely that reports as to the condition of certain muscles are not included. Such a conclusion cannot apply to men like Andre Thomas, Harris, Sherren, Thorburn, Head, et al., whose cases are included. The degree of involvement of the muscles supplied by the collaterals arising near the roots, should be considered in the differential diagnosis of a root lesion. We believe that this is most important, and that a report should state their condition. Many times this information will clear up a seemingly complex problem. Nevertheless, even the reports in this selected series of twenty-four cases, many of them from masters of neurosurgery, show how few times the muscles supplied by these collaterals were definitely recorded as injured or as not injured. It seems highly probable that their condition was ignored in some cases, and that if their condition had been carefully examined, paralyses would have been found more often. In the case of the lower roots the same method was pursued, but here it was more difficult, as there was no case of frankly cut or even injured 7C alone, except Mingazzini's case. The Dejerine-Klumpke type has been seldom checked by operation, except in the four cases where it formed part of a complete paralysis, and no root conclusion could be drawn. The cases of Recaldoni, Pfeiffer, Ransom, Gosset and Quenu, et al., which, while not of cut roots, were at least involvements of the 8C and ID in pathological processes, were also utilized, together with many others, in arriving at what we believe to be at least an approximation of the muscle innervation of the lower roots. The rest was by exclusion, e.g., many cases were reported as injuries of the 5C, 6C and 7C. Not one of these included the paralysis of either the flexor carpi radialis or the flexor longus pollicis. Therefore, we assume that these muscles do not receive their nerve supply from the upper roots. The table of twenty-four cases where the lesion was known and the paralyses were of Erb-Duchenne type is interesting as a diagnostic study. It tends to prove that no distinction can be drawn between lesions of a 5C, a 5C and 6C, or a 6C lesion alone. It shows how little weight is to be placed on the lack of the report of involvement of the root collaterals even in a frankly Erb-Duchenne type. On the other hand, paralyses of the supraspinatus, infraspinatus, deltoid, biceps, brachialis, and brachioradialis may be confidently expected, as shown here, always to occur in this form of paralysis. If the muscles supplied by the root collaterals are paralyzed, it helps the diagnosis, but this table clearly shows that if they are not included, it just as certainly does not rule out a complete lesion of these upper roots. There were twenty-four cases involving the 5C or 6C, where we could localize the injury sufficiently for our purpose. In two of these the 7C was also slightly involved, but not enough to be of much importance, so they are included. One of these had a slight involvement of the triceps, but the other was purely Erb-Duchenne. One other case not included in the twenty-four, Winnen's case, involved a 4C connection to the 5C, at the point of junction, it was said.

Conclusions from Analysis of 710 Reported Cases.

In the series of cases from the literature, beginning with Flaubert in 1827, I found that many were of little importance, since the details were too meager to give us much information as to the methods of production, or even the exact muscle paralyses, and in most of them there was no operation or autopsy by which to check. With all the World War wounds, it would seem that there should have been details in an immense number of cases. Such, unfortunately, was not true. Thousands of cases of peripheral nerve wounds, and some in great detail, were reported, but where the brachial plexus was concerned, the description of the operation in most cases was confined to the wounds, with no detailed account of- work on the plexus itself. These cases were reported by many men. Gosset—fourteen. Moynihan—eleven cases of brachial plexus; four upper, seven lower. Tuffier—280 operations on nerves, but no brachial plexus. Boinet— twenty-five cases of brachial plexus injuries. Tinel—in 639 nerve lesion reports, twenty-seven brachial plexus. Price, Feiss and Terhune —sixty cases of brachial plexus injuries; five complete. Mauclaire— four cases, one axillary, none of them complete. Chiray and Roger, Leclerc, Benisty, Fere, Wiart, et al., reported others. In all, there is a lack of detail, so that we can get very little help from their analyses for our particular purpose. In 710 case reports examined, there were 135 cases in which an operation or autopsy had been done.

Let us analyze the eighty-six supraclavicular injuries in order to avoid the fallacious conclusions which we think other students of the subject have come to. Eighty-six cases of supraclavicular rupture would seem a large proportion of the 135 cases on which operation or autopsy had been done. Doubtless among the remaining 575 which were unchecked, and in which no conclusions could be drawn other than the incidence of paralysis, the proportion of supra- to infra-clavicular injuries might have been the same, but as to that we can only theorize. To determine the cases of actual rupture, we must exclude the following fifty-four cases from the eighty-six supraclavicular injuries.

This leaves only thirty-two cases which represent instances of trauma other than cuts or wounds, where roots were reported as separated either because the two ends were found, or because the traumatic neuroma or scar was of such a nature as to lead the operator to believe that such was the case. Of the cases of Tubby, Tuffier, Bardenheuer, Bristow, Hartley, Banks, Thorburn, Sherren, Kalb, Fisk, Ginsburg, and Thomas, there can be no doubt; we may include also the four cases of avulsion; but of some cases where the report was made of a neuroma or a disintegration of the plexus, or of scar of such a nature as to prevent actual knowledge, there is grave doubt, not of injury, but of an actual rupture of any part of the plexus. The German and Austrian literature has been examined less thoroughly, but including Bernhardt, Wolfler, Winnen, Erb, Kalb, Kramer, Bardenheuer, et al., and excluding all the cut roots or wounds, we have been able to accept only thirty-two cases as definite supraclavicular ruptures! I have found only four proven cases of rupture de la moelle epiniere (avulsion) in all the history of medicine.

Flaubert of Rouen, France, 1827. Reduction of a dislocation by the combined force of eight men. All roots except the 5C ruptured from the spinal marrow and recovered below the clavicle at autopsy. Death on the fourth day. Apert, France. Boy, arm caught in the wheel of a heavy cart; thrown over and over. Autopsy thirty-two and one-half years later. Complete rupture at the spinal cord of the 7C and 8C and the ID. No ascending degeneration of spinal tracts, but the brain area was smaller. Boyer, Toronto, Canada. Birth case. Arm pulled. Rupture more pronounced of the 7C, but 5C-6C-8C and ID are included as injured. Ascending degeneration marked. Brain area involved smaller. The autopsy was made forty-one years after the injury. Charles Frazier, Philadelphia. Man struck by another man falling three stories; 6C, 7C and 8C ruptured, from the spinal marrow. Laminectomy, for intolerable pain, afforded the opportunity to study the exact seat of the injury. Dr. Frazier cut the ID and the 5C sensory roots. In this case there was a suspicion of a fractured vertebra. These are the only four cases I can find in the literature with proven avulsion, but these cases have been responsible for the misconception that the usual locations of injuries to the plexus are within the spinal canal or in the bony gutters proximal to the snubbings on the transverse processes.

Treatment.

Of the non-operative treatment little need be said except that a paralyzed muscle recovers quicker when not in a condition of stretch. One must know and analyze the individual muscle paralyses in order to meet this indication. If one learns to prevent contractures which follow in these paralyzed cases from over-action of muscles not paralyzed, he will have gone a long way in the intelligent treatment of brachial plexus injuries. In cases of complete interruption of the nerve paths, contractures are usually soft and reducible. Only in irritations do we get the hard, irreducible griffs, and these must be prevented by treatment. It is easier than to correct them after they have occurred—in the domain of the median, the accoucheur's hand; in the domain of the ulnar, the claw hand; in the domain of the radial, the wrist drop.

Operation.

Immediate operation is indicated in all cases of penetrating wounds with paralysis involving the branches of the plexus, and in all cases showing the complication of subclavian or axillary arterial injuries. The presumption should be against operating in cases of tension or traction injuries. Were I to select the factor of most importance in determining the question of operation, I should wish to know the manner of the trauma. Lesions of the plexus accompanying dislocation and near dislocation will rarely cause actual rupture, and, therefore, may recover as well without operation as with it. The chance of suture of a ruptured nerve trunk is a forlorn hope. The main indication for operation is the release of pressure from exudate. The resection of a nerve simply because it has a traumatic neuroma or local enlargement is unwarranted;- there must be clear evidence of entire anatomic separation. False neuroma is often hard to distinguish from true, and it will many times interrupt the electric current. See Henri Claude, Dejerine, et al. Lysis, simple longitudinal incision, as taught by Bardenheuer long ago, will give the greater percentage of cures, and resection should be reserved for those cases showing actual separation. The time to operate in secondary cases is of importance. Some say three months, but there are many cases that will not begin to show first motor return in that time, and therefore, if operation were to be the rule, it would be better to operate at once. In expert hands immediate operation is justifiable in any doubtful case, but a surgeon without a knowledge of the local anatomy and general history of the subject might easily do harm. In most cases recovery of function begins within seven or eight months, but many cases have recovered first motion at a later period. In one of my cases, a root lesion due probably to abduction and external rotation during a breast amputation, the first voluntary motion appeared at the end of five and one-half months. Faradic contractility and voluntary motion were noticed on the same day. The best argument for not operating in tension injuries is that actual rupture of the brachial plexus is a rare condition, although temporary palsy is common. The reason is that the stresses do not fall entirely on the brachial plexus as in our theoretical traction apparatus. After the bones and ligaments are separated, stresses are distributed to all the fascial investments of the axilla and neck and so disseminated. Much of the strain, especially when transmitted from below, is taken up by the tendons as evidenced by rupture of the long head of the biceps, so often observed in dislocations and in fractures of the upper end of the humerus. Always, as explained on p. 348, there are small tears and injuries of the fascia about the nerves and vessels, and probably also far removed along the fascial planes. There are petechial hemorrhages and exudate which surround not only the roots but the cords of the plexus in both the supraclavicular and axillary regions. But if rupture occurs, it is at the roots or of terminal branches and not of the plexiform part, which, nevertheless, may be swollen by pressure of the exudate within the fascia surrounding it.

Some Practical Rules in Diagnosis for Use with Chart.

There are many exceptions, but the following points will usually hold good. If in an upper type lesion the brachioradialis is involved, it is usually a lesion of the 5C or 6C roots. If the brachioradialis is not involved in a purely radial or extensor paralysis, often entire at first but later only partial, it is a lesion of the 7C. In such a case the coraco-brachialis, when it comes from the 7C as it often does, would be the only flexor in the upper arm paralyzed and would become the key muscle. It must not be forgotten that in the wrist drop from paralyzed forearm extensors due to lead poisoning, the integrity of the brachioradialis and supinator is almost diagnostic, and this condition must be excluded. If all the extensor muscles of the arm and wrist and fingers, and the brachioradialis and supinator also are paralyzed, it is a peripheral lesion of the radial nerve. If only the deltoid and teres minor of the upper group in addition to the above are involved, it is a trunk and not a root lesion of the intermediate cord or a peripheral lesion of two nerves—radial and axillary. If it is a combined paralysis of the flexors of wrist, thumb and fingers, except the brachioradialis and the pronator teres, and all extensors are involved, it is a lesion of the 7C and 8C and ID. If the lesion is as above but only partial of the triceps, the latis-simus dorsi and partial of all the extensors of the wrist and fingers, with escape of the brachioradialis and extensor carpi radialis longus and brevis, it is a lesion of the 8C and ID roots. If the paralysis is of all the flexors of wrist, thumb and fingers with the exception of the brachioradialis and the pronator teres with escape of the triceps, anconeus and all extensors, it is a lesion of the medial fasciculus. If only the ulnar flexors of the forearm are involved and all the hand muscles except the opponens pollicis and the abductor pollicis brevis and one-half the flexor pollicis brevis, it is a lesion of the ulnar.

If it is a paralysis of the flexors of the wrist and fingers, including the flexor longus pollicis and the pronator teres, with escape of the brachioradialis and supinator, the flexor carpi ulnaris, part of the flexor profundus digitorum and all interossei, it is a peripheral lesion of the median. In this case the abductor pollicis brevis and opponens pollicis and part of the flexor pollicis brevis are also paralyzed, but the hand muscles and adductors of the thumb are not involved. The abductor pollicis longus, it must be remembered (the extensor ossis metacarpi pollicis of the old terminology), is an extensor innervated by the radial. Nevertheless, it may act as an abductor of the thumb. It is not involved in a pure median paralysis. Care must be taken not to confuse a root lesion with a peripheral paralysis which involves two or more nerves. In such a case diagnosis is more often difficult and will tax the knowledge of the observer.

Lesions of the brachial plexus complicated by...

CHAPTER XII

LESIONS OF THE BRACHIAL PLEXUS COMPLICATED BY RUPTURE OF THE SUPRASPINATUS TENDON

INSTANCES will arise in industrial clinics where it will be difficult to differentiate between a direct injury to the supraspinatus tendon and a circumflex paralysis. The most important point I have in mind to accent in this chapter is that one should never be misled by the fact that the patient's deltoid is paralyzed, into thinking that the supraspinatus is undamaged. As a matter of fact, the combination of these two injuries is not infrequent, and the supraspinatus injury remains undetected because the deltoid paralysis seems to be accountable for the fact that the patient cannot raise his arm. I want to present arguments for my conviction that in case a patient has deltoid paralysis and is unable to elevate his arm, we should make an exploratory incision in the bursa to determine whether the supraspinatus tendon is ruptured. This dictum is supported by my observation that if the supraspinatus is intact, i.e., neither ruptured nor paralyzed, the patient can feebly raise his arm in spite of paralysis of the deltoid. I believe that the industrial surgeon who bears this in mind may be able to save his company some trouble and expense. In the previous chapter, Dr. Stevens has made it very clear that most of the postdislocation paralyses are due to local, peripheral injuries of the axillary and musculocutaneous nerves, caused by actual stretching of these nerves by the head of the humerus in the course of its process of dislocation. He has also shown that the suprascapular nerve is subj ect to a considerable amount of variation in its association with the 5th root, so that many times it may escape paralysis, even in cases otherwise of Erb-Duchenne type. Although I have never studied the plexus in regard to the mechanical stresses caused by different forms of injury, as Dr. Stevens has done, I had come to somewhat the same conclusions from clinical study, for I had observed cases in which it was very clear that the site of the lesion was below the nerve supply of the supraspinatus and above that of the deltoid. My observations had brought me to the conclusion that one should always be suspicious that the supraspinatus may be ruptured, rather than paralyzed, in all cases of postdislocation palsies. For the purposes of our particular study concerning the diagnosis of paralysis of the supraspinatus versus rupture of its tendon, either of which combined with deltoid paralysis would cause inability to perform elevation, observation of the conditions of the rhomboids and pectorals is important, as the nerves for the rhomboids and those for the clavicular part of the pectoralis maj or leave the plexus above the circumflex, and also above the suprascapular nerve which supplies the spinati. One can, by observation alone, determine the question of whether the pectorals or rhomboids are paralyzed, but the spinati under cover of the acromion are not readily palpable, although in a perfectly normal shoulder it is easy to feel (through the trapezius) the contraction of the supraspinatus when appropriate effort is made by the patient. When the patient folds his arms across his chest, and then shrugs the shoulders, the contours of the rhomboids appear. However, in cases of injury where the whole shoulder is tender, it is difficult to make sure that the patient is really making an effort to contract his supraspinatus. Thus the positive is pretty positive, but it is difficult to say that you are sure that the supraspinatus is out of commission. Later atrophy of these muscles does not signify paralyses, for it occurs after almost all shoulder injuries. Nevertheless, if the rhomboids are paralyzed and the pectoralis major is partly paralyzed (the clavicular portion), we may assume that the supraspinatus is, also. Since the nerves to the pectorals have an additional supply from the lower cords of the plexus, paralysis of the pectorals need not necessarily be found. It is especially important to observe the clavicular portion of the pectoralis major, for this is probably always supplied by the 5th and 6th roots and perhaps by these roots only. Of course, proof of paralysis of the supraspinatus does not necessarily mean that there has been no rupture, for this might have occurred simultaneously with the same injury which dislocated the humerus and injured the plexus. If the rhomboids are paralyzed it makes it more probable that the supraspinatus is merely paralyzed and is not ruptured, but it does not absolutely exclude rupture. As a matter of fact, in injuries to the plexus from dislocation or near-dislocation, we commonly find only the axillary nerve involved and consequently only the deltoid and teres minor paralyzed. The teres minor is naturally overlooked. Often the biceps and coraco-brachialis and brachialis anticus are simultaneously involved, and not infrequently the supinator longus. Yet all these may be involved without the spinati, for the suprascapular nerve may be almost independent of the 5th root. Of course, testing the activity of these muscles implies cooperation on the part of the patient, who must exert his will to raise the arm in spite of the soreness from the recent injury. Fortunately, in cases of paralysis the sensation of the joint is usually absent owing to the sensory paralysis of the axillary. In a recent severe case in which all the tendons had also been evulsed and the axillary paralyzed, I could move the joint about in any direction, without causing the least pain. Taking all these considerations into account, I should lay it down as a dictum that if, following such injuries, one finds the deltoid paralyzed and the patient is unable to slowly raise his arm with the supraspinatus alone, the probability of a diagnosis of a ruptured supraspinatus tendon is so likely that exploration of the bursa should he done, unless there is a coincident paralysis of the rhomboids or the clavicular portion of the pectoralis major. The following cases are illustrative of some of the points considered :

CASE 30 On Sept. 3rd, 1912, I operated on a man of 47, who had a circumflex paralysis which followed a dislocation two months previously. I was in doubt as to whether the supraspinatus was torn. The bursa, on exploration, proved to be normal, and the supraspinatus tendon was intact. It occurred to me that as the patient must wait for his incision to heal, he might as well wait in abduction and external rotation and give his deltoid a rest, on the general principle that all paralyzed muscles recover their tone better if they are maintained in a relaxed position. After a few days this man began to be very complimentary about the success of my wonderful operation, and said he could move his arm better than he had since the accident. In two weeks the power of the deltoid had almost entirely returned and the patient was greatly pleased.

Of course, my operation being simply a small incision through the deltoid fibers and the roof of the bursa had nothing to do with the matter. His improvement was due to rest in elevation. Since this experience I have treated deltoid paralysis in the same way, sometimes without exploratory incision, with invariably gratifying results. Most of the "improvement takes place in the first two weeks, so that the patient is usually very cooperative, for he realizes his progress. Part of the rapid progress in the first two weeks is due to readjustment of the swelling and edema in the joint and bursa. If the patient has been up and about for several months and the shoulder has "dropped," the space between the acromion and the head of the bone must become filled with fluid or with edema of the joint and bursal tissue. The elevated position gives this edema an opportunity to be absorbed, the circulation becomes more active and the muscle fibers, not being stretched, soon respond. It may be months, however, before the muscle becomes fully developed again. The principle of relaxing paralyzed muscles is well accepted, but so far as traumatic circumflex lesions are concerned, it has certainly not become an appreciated routine practice, in this community at least. Previous to the above experience I had never seen it applied to a case of circumflex paralysis, and even since the above experience, my own cases have been the only ones I have observed, although I have no doubt the principle is constantly gaining in the extent of its recognition and perhaps practice. (See p, 482 on relaxed capsule.) It is well recognized now in the treatment of infantile paralysis. I feel sure that there are many industrial cases of axillary and other paralyses, for which our insurance companies might well pay for the use of a hospital bed for the sate of the time which the patient would gain in returning to work, As a rule, the exploration of the bursa should be done with local anaesthesia at the same time, for it is a trivial operation and causes no delay. The wound is healed long before the use of the arm returns. If a rupture of the supraspinatus is found and repaired, permanent disability may be prevented.

CASE 49. On Dec. 13th, 1913, I operated on a man of 64, who had an axillary paralysis following a dislocation three weeks previously. On this case I made a preoperative diagnosis of a rupture of the supraspinatus as well as of deltoid paralysis. .Exploration showed that not only the supraspinatus but the infraspinatus, teres minor and long head of the biceps were evulsed. The destruction was too great to make any suture possible, so the wound was closed and the arm put up in abduction. Within two weeks there was a fair return of the power of the deltoid and months later the deltoid even became hypertrophied, but the full power of abduction never returned. I was called to see this patient ten years later, on April 23rd, 1923, because he had an ununited .fracture of the neck of the femur. The deltoid of the injured arm was very large and powerful, but he still could not abduct the arm when standing.,

Undoubtedly if this patient had been operated on immediately after the accident, the capsule and short rotators could have been sutured in place. Increased experience in these cases makes me feel confident that at the present time I could suture the tendons in a similar case even after three weeks, but the longer the period of time that elapses between the injury and the operation, the more retracted the tissues and the greater the difficulty. Quite recently I had to abandon suture in a case of seven months' duration in which both spinati and the teres minor had been evulsed.

A more encouraging instance was Case 107, a strong woman of 58, who had been employed as a cook. On July 18th, 1927, she fell on the floor of her kitchen and dislocated her left shoulder. She was taken to an emergency hospital, and the dislocation was reduced. The shoulder remained sore, almost useless, and did not progress. After seven months she was referred to the office of the insurer and an X-ray was taken showing that the dislocation had been reduced. Next day, Feb. 16th, 1928, she was referred to me. I reported as follows: "Examination of the left shoulder reveals that there is a sensory and muscular paralysis of the circumflex and little power in the biceps. The head of the humerus falls away from the glenoid, as is usual in these cases of circumflex paralysis. "I would recommend that the patient enter a hospital and have the arm retained comfortably in abduction. I have had a number of cases which responded promptly to this treatment, probably because the nerve trunks are put at rest and their circulation therefore made better for repair. Also the drag of the weight of the arm is taken from the deltoid and its fibers have a chance to recover their tone. "With this treatment the prognosis is good. Two weeks in this position with occasional appropriate exercises produces marked improvement. If it does not, the prognosis is poor, but I have never yet failed to see demonstrable improvement recognized by the patient in this time."

*        *        *        *

"In my report of Feb. 16th concerning this patient, I stated that Dr. reported to me by telephone that the films showed normal bones. I have since called at his office and inspected the films. They are not wholly negative for they show, faintly, crumbs of bone in a position suggesting that they were torn off the greater tuberosity by the supra-spinatus tendon. The films were not clear, for the patient moved, but the suggestion is strong that the damage is not only to the brachial plexus but that a certain amount of evulsion of the superficial part of the tuberosity occurred at the accident. "At your suggestion, the patient entered the Trumbull Hospital on Feb. 19th to be under my care, and I put her to bed with the arm held loosely in abduction, so far as it was possible. I have seen her yesterday and today. She is not very cooperative and I find her a hard patient to handle, as she is rather stupid about getting the idea of keeping her arm up and when the nurse's back is turned, gets it down again. "There is spasm of the groups of muscles on each side of the axilla and this tends to make her arm return to her side, as there is no opposition in the deltoid. It appears to abduct fairly well up to about half the normal distance and then there is a block, as if the above-mentioned fragments impeded further progress. It may be that there is some callous formation about them. However, I shall try for a few days longer to cope with her and endeavor to carry out the treatment. Unless she becomes more cooperative I fear that I may have to discharge her, for I do not think she would be willing to have an exploration of the bursa to see if the tendon is evulsed with a bit of tuberosity attached."

*       *        *       *

"I reported on this patient last on Feb. 21st. Since then she has remained in bed at the Trumbull Hospital with the arm in abduction. I finally succeeded in getting it into her head that she should help, and after the first week she, herself, could notice the daily improvement and has been more enthusiastic with her cooperation. It took about five days to overcome the spasm of the latissimus and pectorals, etc., but when they once yielded the arm could be comfortably retained in a semi-abducted and externally rotated position without other apparatus than a bandage, about her wrist tied to the head of the bed, to remind her. "Her progress has been good, and there is distinct improvement in sensation and muscular activity in the deltoid region. I give her exercises in the stooping posture and she can do them quite freely, the arcs of motion being normal and the pendulum movement taking the place of the muscular power of the deltoid. There is still no real power in the latter, but I am confident I can feel a beginning contraction of the fibers, especially when a counter effort is being made. "In my opinion it is well worth while to keep this patient in the hospital as long as she is clearly improving and she is willing to stay. The result of the two weeks makes me confident that progress will continue and the result will be good. I wish I could also explore the bursa to see whether the tendon is torn or whether the atrophy and loss of power in the supraspinatus is wholly from the nerve paralysis. I suspect the tendon is torn, but as it is masked by the paralysis, I am not certain. "Would Dr. - - approve of my trying to persuade the patient to have this done? The bursa and tendon could be inspected through a very small incision, not over a half-inch long. It would save the patient much time to have this done now rather than to wait until the paralysis would disappear. Then, too, if the tendon is torn, it would be easier and less painful to mend it while the paralysis is present than after the power of the deltoid has returned. If I were in her place I would beg to have this done, but I can see that she is not a patient to be easily persuaded."

Consent being obtained, the operation was done.

"Operation on Mrs. B.j March 13, 1928—ether anaesthesia. An incision was made (not over .one-half inch long) over the bursa and the new instrument I have had made was introduced. It is a modified nasal speculum and worked perfectly. On incising the bursa, free fluid escaped. The diagnosis of rupture of the supraspinatus and infraspinatus could be made and the extent of the rupture determined. The incision was then enlarged into the routine bursal incision, about two inches in length. "Considerable difficulty was encountered in pulling forward the retracted ends of the tendons. The biceps tendon was intact and covered by the edge of the capsule and the tendon of the subscapularis. I eventually succeeded in drawing the tendons together, and, as no stub was left on part of the tuberosity, I drilled a hole and passed a braided silk ligature through it and the heaviest portion of the supraspinatus tendon. I also curetted lightly the raw bone on the tuberosity to stimulate attachment of the tendon. Some crumbs of bone were found in the retracted portion. One bit was removed and another sewed to the tuberosity with the tendon still attached. The muscle was closed with catgut, but the bursa was not sutured. A folded pillow was placed in the axilla. "N. B. In this case an unusually accurate closure was made, perhaps because the retracted muscles were paralyzed and yielded more when pulled forward. A good result should be obtained if the wound heals normally, and, as I fully expect, the muscles regain their power."

The wound healed well and the muscles redeveloped. The patient has now a good strong arm with considerable power in abduction. I saw her last on Nov. 10th, 1930. She still complains of soreness and pain and weakness after using it, although it is a year and a half since the operation. She still receives compensation and feels unable to go to work. The X-ray shows a defect in the tuberosity where the small fragments were removed. The patient admits great improvement, but it must be put down as one of those cases of which the surgeon is proud, but in which the patient is not wholly satisfied. Of course the question of compensation enters the problem. If I could obtain as good a result as this, operating seven months after the accident, it seems to me very convincing that similar operations, done immediately, would be very successful. I have on two occasions explored the bursa in other cases of axillary paralysis and found evidence of partial rupture of the tendon not extensive enough to demand suture. The wound did not in any way interfere with the convalescence, and the power of the deltoids returned satisfactorily. It may well be asked how we are to know immediately after a shoulder dislocation has been reduced, whether the deltoid is paralyzed or whether there is crepitus in the joint from fragments of chipped tuberosities? Do I recommend churning the joint about and risking redislocation and further traumatization ? Might not a deltoid paralysis disappear within a few days? Would I make it a rule to explore every bursa after every dislocation because the supra-spinatus may be ruptured? Let me say emphatically that I believe little if any harm could be done by permitting the patient to move his arm immediately after the reduction. Uncomplicated dislocation cannot recur unless the arm is abducted and externally rotated, but to test whether paralysis is present in the deltoid, it is not even necessary to manipulate the arm. One can hold the elbow at the side and ask the patient to make an effort to abduct. If the deltoid can be felt to contract, that settles the question. This test should be applied both before and after reduction, and can do no possible harm. In case of doubt, I should not hesitate to move the elbow away from the side in internal rotation, and then to let the patient hold it there if he can. Unless the arm is externally rotated or carried to the pivotal position, it cannot dislocate. A case that readily redislocated should certainly be explored. Electrical testing of the muscle for the reaction of degeneration (i.e., lack of faradic response and persistence of a slow galvanic response) is of no help in the first ten days, for it takes about this time for degeneration to occur. Even later on, the simple test of palpation of the muscle during voluntary effort seems to me nearly as reliable and more practical. Chips of bone from the facets should be found by the X-ray either before or after the reduction. In at least two of the above cases they were present, as shown by the X-ray and confirmed by operation. Their very presence indicates that exploration is advisable, unless the fragment is large enough to suggest that the base of the bursa is not torn as discussed under fractures of the tuberosity. The following case will be reported in some detail, not only because it is an illustration of the point we are at present discussing of the desirability of exploration of the bursa in cases of circumflex paralysis following dislocation, but because it illustrates many other points spoken of elsewhere in this book.

CASE 115 On Sept. 10th, 1928, a man of 48 was referred to me by an industrial insurance company. He was a well-built, wiry individual who had never been sick or had any serious accidents, although for most of his life he had been employed as an adzeman in a building and wrecking company, a rather hazardous occupation. On August 21st, 1928, he fell from a second-story roof, and as he fell, caught the edge of the roof of a shed on the first story with his bent and half-abducted right arm, thus sustaining him for a moment while the full weight of his descending body, in an almost upright position, plunged downward. A tremendous force was therefore brought to bear just at the shoulder joint. He was taken to Hospital A, one of our best institutions, where X-rays demonstrated a subcoracoid dislocation. Ether was given, the dislocation reduced and the patient sent home with his arm bandaged. He was attended by a doctor at his home, who readjusted the bandages a few times. On August 29th, this doctor, feeling that there was something still wrong with the shoulder, called the medical clinic of the insurer and requested that they take over the case. On August 30th, a representative of the medical staff of the insurer called to see the patient at his home and found "arm markedly swollen, shoulder black and blue, the entire arm edematous. The appearance of the shoulder at that time looked as though the joint were still out of place. He accomplished none of the usual motions of the shoulder, etc." The doctor then took the patient to Hospital B, where more X-rays were taken, which "conclusively prove that the head of the humerus is in proper relation to the glenoid." The man received palliative treatment at this hospital for a few days, during which the swelling subsided a little, and he then attended the insurer's clinic, where I saw him on Sept. 10th, and found that his deltoid was paralyzed, that he had "dropping shoulder" and fluid in the joint. I advised having him enter another hospital to be treated with his arm in abduction for a week or two. He entered Hospital C, Sept. 11th, and I took charge of him. On Sept. 17th, after having reviewed the X-rays taken at all three hospitals, I reported to the insurer's medical department: "At the first visit I put him up in as near abduction (elevation) as I could, and on each occasion since I have managed to get the arm into still greater abduction. The patient has much improved. The deltoid fibers have contracted so that they are no longer flaccid, and I believe that I can feel a little contraction in them, particularly in the anterior portion. "When the arm is almost completely abducted (elevated), it is quite evident from my examination and from the patient's feeling that the head of the humerus is in approximation with the glenoid, but when the patient stands or sits up, letting the arm fall, the head of the bone is thrust forward and downward, presenting the peculiar resemblance to disloca tion which had been noticed by Dr. - -. On thinking the case over I have become satisfied that this patient has five different lesions. "First, he had a rupture of the axillary vein (P. S. I think now that this was probably the axillary artery itself.), or one of the other large veins, causing a hemorrhage which infiltrated the whole upper arm and upper part of the forearm, and also the subcutaneous tissue on the adjacent side of the chest. "Second, he had a direct trauma to the circumflex nerve, or to the cord higher in the neck, causing paralysis of his deltoid. "Third, a fracture of the greater tuberosity, which resulted practically in a rupture of the supraspinatus, because the fragment of bone (P. S. A tiny one.) has been left down under the deltoid, and the supraspinatus has retracted under the acromion. "Fourth, a dislocation of the humerus which allowed the head of the bone to be pushed down into the axilla, while at the same time the fragment of the tuberosity was pushed below under the deltoid by the acromion. "Fifth, it is possible also that the long head of the biceps is torn; certainly there is something which tends to slip in between the head of the bone and the glenoid as soon as the arm is adducted. " In my opinion, it would be best for this patient, to do an exploratory operation under local anaesthesia, and open the bursa to see whether the supraspinatus tendon is torn, and if so, whether it can be replaced. It probably would be best to leave the displaced fragment to absorb." Soon after this I did operate. "A routine bursal incision was made, and after opening the roof of the bursa, the articular cartilages and joint cavity were directly visible. There was no sign of the torn supraspinatus, which had retracted. The incision was lengthened downwards and upwards, so that it extended from the coraco-acromial ligament downward about two inches; still none of the short rotators were seen. A sheet of tissue resembling joint capsule was cut through in making the incision, and this was not true joint capsule, but light scar tissue, which had formed beneath the deltoid. A large, bare area of bone was exposed at the position of the greater tuberosity. The bicipital groove was visible, but the biceps tendon was not to be seen. The subscapularis was detached from the lesser tuberosity. "By pulling down upon the arm a space could be made between the head of the humerus and the glenoid fossa. The finger could be passed completely around the head of the humerus without encountering any attachment of the short rotators. The joint cavity was washed out with warm saline, and the glenoid inspected. Some soft tissue (P. S. The musculo-tendinous cuff.) covered over the glenoid, and the normal glenoid fossa could be palpated beneath it. This tissue was in all probability composed of the short rotators. It was thought inadvisable, if not impossible, to pull out and readjust these rotators. At no time during the manipulation, although only the skin was anaesthetized, did the patient complain of pain. "This case has been remarkable in many respects. Nothing was accomplished by the operation except to determine the extent of the injury. When the wound was opened it was quite apparent that the nerve supply of all the tissues involved beneath the skin was destroyed. For instance, I could put my finger in between the glenoid and the head of the humerus, and feel all around the head without giving the patient any apparent pain. I could raise the arm in abduction and internal rotation, and move it about at will without starting any pain or any spasm. All the short rotators were evulsed from the bone, and the head could easily have been pushed out through the wound as far as the surgical neck, as we do in excision of the head of the humerus. When the joint was apparently reduced the short rotators and capsule lay across the glenoid in such a way that the head rested on them; therefore, accurately speaking, the dislocation was entirely unreduced, and yet only a film of tissue one-eighth to one-quarter inch thick separated the articular surface of the humerus from the articular surface of the glenoid. The condition was similar to that represented in Fig. 58, except that the fragments of bone were much smaller."

I feel very certain that the force which caused this injury resulted in pushing the head of the humerus straight down into the axilla, with such violence that the whole capsule together with the short rotators and biceps was evulsed entirely, probably in one piece from the attachments on the tuberosities. The infraspinatus and teres minor were torn as well as the subscapularis and biceps tendon. As the head of the bone descended into the axilla, it ruptured an axillary vessel, or probably a large artery. At the same time a chip from the greater tuberosity was pushed by the acromion into the space beneath the deltoid, and when the dislocation reduced, this chip still stayed down there as shown in the X-ray. It was the same force and the same leverage which usually results in fracture or dislocation, but in this case the bone, except for a bit of facet, held, and all the soft parts gave way. The problem of surgical treatment after I had exposed the condition was too difficult to solve. I could have cleaned out the glenoid, but if I had done so, I could not have gathered the ends of the short rotators and resutured them to the tuberosity, even by drilling holes in it, for the tuberosity itself was too much damaged. The only operation that I can conceive of doing in this case, and this might still be done, would be to use a sabre-cut incision, dissect out the short rotators separately, clean out the glenoid, pass a strip of fascia lata through the head of the humerus, and attach the short rotators to it. (Fig. 52.) I did not feel that I could perform such a difficult operation under the circumstances. It seemed to me better to leave the arm in the condition that it was, hoping that a false joint would develop which would permit him to use the other motions of his arm without abduction. The patient was put to bed with the arm in an abducted position, and I let him remain in this position until the wound was healed and the power of the deltoid had returned. This resulted in the patient's feeling much encouraged, so that he would not consent to having another operation performed, feeling that eventually he would recover the remaining use of his arm. Finally, believing that he could get a more useful arm, I urged him to have an arthrodesis performed, but he would not consent to this and insisted on returning to his birthplace in Newfoundland, where he could live cheaply on his savings. He settled with the insurer and left. On May 23, 1932, he wrote me that there was no improvement but that there was no sign of axillary aneurysm. At any time he could have an arthrodesis done which would give him a painless, useful arm, although with the arc of motion limited fifty per cent. I should have done this operation in the first place. He may still develop an axillary aneurysm.

Among the instructive things about this case were: That there was no anaesthetic skin area larger than a fifty cent piece, and this was down at about the middle of the deltoid. Yet when the skin was anaesthetized, the whole joint could be manipulated without any feeling on the part of the patient. This is to be explained by the fact that the nerve supply of the joint is from the deep branches of the axillary and also largely from the suprascapular, which was removed from the field by its retraction with the capsule over the glenoid. The seat of incision in the skin was supplied by the cervical plexus, which was not injured. Although this case showed all the common complications of dislocation, I feel that many cases have the main one of rupture of the supraspinatus. As a matter of fact, after satisfactory reduction of an uncomplicated dislocation, the patient does not need immobilization, and a certain amount of active and passive motion is desirable, and neither painful nor harmful in any way. I believe that fixation in the sling position for ten days or two weeks after dislocation is actually harmful, and that routine exploration of the bursa in every case would be preferable. Yet I do not advocate routine exploration, although I believe it should be done in all cases of deltoid paralysis that cannot voluntarily abduct the arm, in all cases where evulsed facets are demonstrated by the X-ray, and in all cases where the surgeon is uneasy as to whether the dislocation is properly reduced. Any such rule may have exceptions. I have never known or seen a case of deltoid paralysis which completely disappeared in a few weeks, let alone in a few days; it always takes weeks or months for the complete recovery. If the patient is to be laid up some time, why not make the exploratory incision at once, and if no rupture is found, get the benefit of the rest with the arm elevated? The reader may question my dictum that when the supraspinatus is not injured the patient can abduct his arm by the use of this muscle alone. My evidence comprises a number of cases personally observed and a few cases from the literature. Perhaps the most convincing case which I can report was that of a young Italian laborer who had absolutely no deltoid muscle, but who could elevate his arm, not only easily, but with great power, with an hypertrophied supraspinatus. This patient had been seen by a number of other doctors, and the question of the existence of a bone sarcoma had been brought up, because the shoulder was so misshapen that the appearance suggested a tumor just above the scapula. Some one also had made the diagnosis of progressive muscular atrophy. The following is a copy of my report to the insurance company.

Sept. 30, 1926. This patient is a rather small Italian laborer, age 26, whose face is distorted and scarred. He claims that this condition of his face is due to wounds received during the war in Italy. Otherwise than that, he says he has been well, until April 29, 1926, when he was climbing a ladder, holding a pail in one hand, and a rung broke and he fell. He does not remember how he fell, nor on what he landed, but the records submitted with him state that it was three stories. He was taken to a hospital and treated for fracture of two of the left lower ribs and an abrasion of the hand. He was soon after discharged from the hospital, but within a week after this developed pneumonia and had to return. With the request for the examination of this patient you were kind enough to submit a folder containing a portion of his record. This folder does not contain records of his case prior to August, so that I am unable to ascertain either from the record or from the patient just what was the condition of his shoulder during the first few weeks after the accident. It appears from your record that the tumor in the supraclavicular fossa was not noticed until August. I can only give my impression from the examination at this date without a real history. The X-ray, which you also kindly sent, rules out any lesion of the shoulder bones and indicates that the tumor is subfascial. Examination: The scars on the left forehead and cheek indicate that he had an old wound resulting in entire disappearance of the left masseter muscle and perhaps some loss of bone. The bulging side of his face is the normal one, and the disparity of the two sides is due to the contrast of the normal with the atrophied side, rather than vice versa, as one would guess at first sight. The deformity of the face is so great as to suggest a "facial hemiatrophy." One must consider the possibility that this was a congenital condition and that it was not related to the wound and scars. The examination of the shoulder discloses a most unusual condition, namely, a complete atrophy of the whole deltoid muscle without the involvement of the neighboring muscles, except as shown by a mild fibrillation. It is very interesting that the patient can perform complete abduction of the arm with the supraspinatus alone without the deltoid. He does this so easily that it suggests that the present condition has existed much longer than since last April. Another point which suggests this conclusion is that the patient's arms are covered with tattoo marks, which are not misshapen in any way, as they would be if they had been done before the atrophy of the deltoid occurred. He states that these tattooings were done in 1923. Referring to the tumor in the supraclavicular fossa, there is certainly a hard mass at this point, but I do not feel sure that it is a tumor in the sense of a new growth. It is subfascial, firmly fixed to the upper ribs, but not to the scapula. It is not tender to any great extent. It is immovable, smooth, hard and not elastic. It is covered more or less with muscle and is difficult to describe accurately. I am not sure that it is not a peculiar curvature of the ribs due to the unusual shoulder condition which may have existed since infancy. There is one other finding which may have some bearing, and that is a partial if not complete paralysis of the left serratus magnus, causing a slight angel wing appearance of the posterior edge of the scapula. Discussion: This case has puzzled me a great deal, but I have come to the conclusion that it is an instance of an old infantile paralysis, and that the recent injury in April had little or nothing to do with the condition of the shoulder. I do not think that the diagnosis of progressive muscular atrophy is correct, although I would admit that it might be if good proof could be obtained that the patient possessed a deltoid within the last few years. The extraordinary development of the supraspinatus muscle is very much against this hypothesis. I am even inclined to doubt that the facial condition is due, as the patient says, to wounds received in the army. I would be much more inclined to think that the man "got by" in a physical examination and entered the army, as many of our own men did, without a thorough looking over. On the whole, I am of the opinion that no treatment will do this patient any good, and that he, himself, knows that he is trying to put over conditions which he previously had knowledge of, as results of the recent accident. I can feel slight irregularities on some of the lower ribs which suggest that they have been fractured, but nothing that I would be willing to declare were evidence of fracture. The patient had noticed no tumor in the supraclavicular fossa, nor does he complain of symptoms from it, nor does the X-ray show any definite indication of a true tumor. From the patient's point of view, exploration of this tumor might be worth while. The possibility of its being a neurogenic fibroma or sarcoma has to be considered, but I am inclined to think that exploration would show it to be a compensatory deformity of the rib, due to the use of the arm without the deltoid since childhood. I cannot, however, believe that it is a part of the duty of any insurance company to have this done for the patient. Notice the similarity of development of the two hands and forearms. This is consistent with a deformity of the shoulder in infancy, and not with one beginning April last.

I finally convinced the neurologist by an X-ray comparing the two scapulae which showed that of the affected side to be much smaller than the other, and he withdrew his tentative diagnosis of progressive muscular atrophy. The upper three ribs were bowed outward and gave the appearance of tumor. As a matter of fact, the cause of the atrophy of the deltoid in this case is unimportant for my argument. Whatever its cause, the deltoid muscle was gone, and the supraspinatus was observed to function as a powerful abductor. I have from time to time seen other cases in which the supraspinatus was able to perform abduction in spite of a paralyzed deltoid. Quite recently I saw such an instance in a patient of my own, who was riding a motor cycle and collided with a truck. Besides minor cuts and bruises, he had a definite paralysis of the deltoid, but in spite of this he could elevate the arm to an upright position. After several months the deltoid returned to normal, and a year after the injury was as well as ever. In this case, also, the sensory area corresponding to the circumflex distribution was no larger than a half dollar, if we may judge from the skin anaesthesia.

SENSORY LESIONS OF THE SHOULDER

Before leaving the subject of injuries to the brachial plexus, something should be said in regard to whether the sensory roots may be affected without involvement of the motor roots. So far as I know, such cases have not been described, and it is even very rare to find any mention of sensory involvement in the descriptions of cases of motor paralysis of the upper type. In the lower type the skin anaesthesia is in the hand and arm, and not in the shoulder. Probably most neurologists would at once suspect hysteria, if anaesthesia over the shoulder without motor involvement followed an accident, yet I am convinced that such symptoms may arise from bona fide organic causes. The fact that anaesthesia so seldom accompanies motor injuries is an additional argument to support Dr. Stevens' contention that ruptures of the plexus are usually external to the snubbing on the transverse processes. If the injury extended within the spinal canal, the posterior sensory roots should be involved. But may these roots, or their posterior sensory branches, not be involved separately? I am inclined to think that they may, perhaps not within the spinal canal, but after they have emerged from the bone and are perforating the deep layers of fascia on their way to the skin. In certain cases I have found areas of extensive paresthesia in the regions supplied by the posterior branches of the cervical and brachial plexus over the dorsum of the shoulder. There was localized anaesthesia or an intense hyperesthesia in this area. There was no definite muscular paralysis, yet the individuals were incapacitated on account of pain and soreness on the use of the shoulder. I am not able to state exactly what this condition is, but I am inclined to interpret it as an injury to the posterior sensory roots of the brachial plexus at a point between the dura and the place where they penetrate the deep fascia in the back near the vertebrae. It seems possible that a tearing of the heavy fascia at this point might stretch or pull off a number of these roots, without doing any damage to the motor roots.

Case 1. W. J. A muscular, strongly-built teamster of 37. On Feb. 7 th, 1929, he jumped up several inches from the floor to pull down a lever in a factory and felt a sudden, sharp pain in his neck and was immediately unable to use his left arm. He saw many doctors and had various forms of palliative therapy. I saw him on July 31st, 1929. He stated that there had been no great change in his condition since he was injured.

Examination showed a powerfully built man in apparent health, who held his head and shoulders in a peculiar manner as if dreading any twist or turn of his body. The posterior muscles of the right side of the neck were very prominent, as contrasted to the depressed condition of those on the left. It seemed to me that this lack of symmetry was due to lack of function of some of the left muscles rather than to spasm of the ones on the right, yet the trapezius could be contracted to hunch the shoulders so that the muscles affected, if any, must be the deep group. No special muscle could be identified as paralyzed. I found an area of intense hyperesthesia to light stroking of the skin of the back of the left side of the neck and chest, corresponding to the distribution of the dorsal branches of the cervical and brachial plexus; yet this area was not sensitive to a pin prick.

A second examination was made Feb. 3rd, 1930. There had been slight improvement, but the condition was essentially the same. Motions of the joint and the development of the muscles were normal.

FIGURE 66. AREAS OF SENSORY DISTURBANCE UNACCOMPANIED BY PARALYSIS The disturbed sensation area in Case 1 involved only the back. In Case 2 it involved the back and shoulder. In Case 8 it involved the back, shoulder and outer and anterior portions of the arm. There was no definite evidence of paralysis of any of the muscles, although the writer could not absolutely determine that there were no paralyses of the deep muscles of the neck. The curved posterior margin on the back was identical in all three cases.

Case 2. J. M. A red-haired, strong-looking Irish laborer of 24 was struck unexpectedly from behind by a huge roll of leather, which was immediately followed by another roll which also struck him. He was knocked to the floor unconscious and taken to a hospital. He sustained injuries to the hip and ribs and to his left shoulder. I saw him on June 27th, 1929, nearly six months after the accident. He stated that the symptoms had been the same ever since the accident, so far as the shoulder was concerned. He had completely recovered from his other injuries. Examination showed that the left shoulder joint and the muscles about the shoulder were normal. He, too, held his neck in the same peculiar position. There was tenderness about the upper cords of the plexus and pain in the rhomboid region. There was anaesthesia to a pin prick over the whole area supplied by the posterior branches of the left cervical plexus, and by the circumflex and posterior branches of the upper five or six thoracic nerves. The posterior outline of the anaesthesia was similar to the outline in Case No. I, but the area of anaesthesia extended over the circumflex area also. I saw this patient again on December 10th, 1929, and the anaesthesia remained the same. This case was considered to be hysterical by a neurologist, but I feel sure it was not.

Case 3. J. D. On October 14, 1929, a longshoreman of 42 was struck on his left shoulder and left side of his neck by a heavy load swinging on a hoist, and was knocked about fifteen feet. I saw him on March 3rd, 1930, nearly six months later. He stated that his condition had been about the same since the accident. He complained about inability to use his left arm without pain. Examination showed a similar paresthesia to that in Case No. 2, except that in addition to the area involved in Case No. 2, the areas supplied by the median and radial nerves in the forearm and hand were also involved. In this case also there was no paralysis or wasting of the muscles in the arm. They were, however, somewhat weak. There was some delayed sensation in the ulnar supply of the little and ring fingers and ulnar region of the forearm. This would seem unlikely in an hysterical case.

Several other similar cases have come under my observation. They are mentioned in this connection because my belief is that this group of cases constitute a real clinical entity, although the peculiar character of their sensory symptoms, without motor symptoms, suggests hysteria. I am sure that similar cases will be met by other surgeons who see industrial cases, and feel that they merit study and discussion. In the three cases the posterior outline of the paresthesia was the same. Unfortunately I have no treatment to suggest. Within six months after writing the above, I saw two more very similar cases of this puzzling type. They would have escaped my notice had I not had the previous experience. I feel that it is important to test the dorsal region of the neck and shoulder for anass-thesia or hyperesthesia, particularly in strong, muscular individuals, such as all these men were. They were not at all of hysterical type, and yet the only findings to account for their great and continued complaint of pain were these unusual areas of paresthesia. There is a possibility that these areas of paresthesia represent overuse of physiotherapy. All these cases had had prolonged treatment. I am convinced that "Baking," whether by hot water bottles or electric pads and lamps, may produce similar disorders of the nerve terminations in the skin. However, these patients all stated that their symptoms were the same or worse soon after their accidents. Although I cannot explain the exact mechanism of the injury in the cases described, I am loath to believe that previously sound, muscular men, after admittedly serious injuries, should succumb to hysteria and exhibit its symptoms by skin paresthesia in this region with a definite posterior border arching away from the spine. The fact that the posterior branches of so many different nerves are involved suggests a long, vertical rent in the deep fascia of the neck.

REFERENCES

I know of no writings on the subjects considered in this chapter.