Clinical repercussions of Martin-Gruber anastomosis: anatomical study.

OBJECTIVE: The main objective of this study was to describe Martin-Gruber anastomosis anatomically and to recognize its clinical repercussions.
METHOD: 100 forearms of 50 adult cadavers were dissected in an anatomy laboratory. The dissection was performed by means of a midline incision along the entire forearm and the lower third of the upper arm. Two flaps including skin and subcutaneous tissue were folded back on the radial and ulnar sides, respectively.
RESULTS: Nerve communication between the median and ulnar nerves in the forearm (Martin-Gruber anastomosis) was found in 27 forearms. The anastomosis was classified into six types: type I: anastomosis between the anterior interosseous nerve and the ulnar nerve (n = 9); type II: anastomosis between the anterior interosseous nerve and the ulnar nerve at two points (double anastomosis) (n = 2); type III: anastomosis between the median nerve and the ulnar nerve (n = 4); type IV: anastomosis between branches of the median nerve and ulnar nerve heading toward the flexor digitorum profundus muscle of the fingers; these fascicles form a loop with distal convexity (n = 5); type V: intramuscular anastomosis (n = 5); and type VI: anastomosis between a branch of the median nerve to the flexor digitorum superficialis muscle and the ulnar nerve (n = 2).
CONCLUSION: Knowledge of the anatomical variations relating to the innervation of the hand has great importance, especially with regard to physical examination, diagnosis, prognosis and surgical treatment. If these variations are not given due regard, errors and other consequences will be inevitable.

Introduction

The anatomy of the upper limbs is extremely complex. In addition to the complexity of the format of the brachial plexus and the existence of strategic anatomical zones such as the epitrochlear-olecranon groove (ulnar canal), carpal canal and Guyon canal, anomalous nerve branches can be found. These may form anastomoses in particular sites that have clinical and functional importance.

The innervation pattern that is most accepted in relation to the intrinsic muscles that act in relation to fine movements of the hand is that the opposing short abductor muscle, the superficial portion of the short flexor muscle and the first and second lumbrical muscles receive innervation from the median nerve. The deep portion of the short flexor muscle, the adductor muscle of the thumb, the muscles of the hypothenar region, the dorsal and palmar interosseous muscles and the fourth and fifth lumbrical muscles are innervated by the ulnar nerve. The great variety of clinical conditions observed in cases of isolated lesions of the median or ulnar nerve led us to investigate this.

Clinically, variations of the innervation of these small muscles are very important, in that even if the median or ulnar nerve is completely torn, some of these muscles may not necessarily be paralyzed. This could lead to the wrong conclusion that the nerve had not become completely torn.

The studies that we consulted showed many divergences, and these became more marked when the authors used different investigative methods. The methods used included electroneuromyography, selective anesthetic blockage of nerves and anatomical dissection. Clinical reports and electroneuromyographic studies have suggested that these muscles may receive double innervation from the median and ulnar nerves. Highet analyzed 45 patients with completely torn median or ulnar nerves and found that the short flexor muscle of the thumb was totally paralyzed in only five of them. Forrest drew attention to this and stated that communication between the median and ulnar nerves in the hand might be more frequent than is imagined and that the muscles located close to this communication (nerve anastomosis) might be receiving innervation from both nerves (double innervation).

Reports on clinical cases have often suggested that double innervation of these muscles exists. Nerve communication (anatomical variations) between the median and ulnar nerves may occur in the forearm (Martin-Gruber anastomosis), between the thenar motor branch of the median nerve and the deep motor branch of the ulnar nerve in the palm of the hand (Cannieu and Riché anastomosis) or between the sensory branches of the two nerves, also in the palm of the hand (Berretini anastomosis or superficial sensory communicating branch). Thus, the variety of clinical conditions observed in cases of isolated lesions of the median and ulnar nerves is discordant with the classical innervation pattern of these muscles that is suggested by treatises on anatomy. Better knowledge of the anatomical variations in the innervation of these muscles is important with regard to diagnosing nerve injuries.

Nerve communication between the median and ulnar nerves in the forearm, which is known as Martin-Gruber anastomosis, causes transfer of nerve fascicles from the median nerve to the ulnar nerve. This gives rise to alteration of the normal anatomical pattern of the motor and sensory innervation of the hand. It was first described by Martin, who considered the possibility of communication between these nerves in the forearm. Around 100 years later, in 1870, Gruber dissected 250 forearms and found communication between these two nerves in 38 of them.

It has been demonstrated in the literature that the incidence of this nerve communication varies when different investigative methods are compared, i.e. according to whether electroneuromyography, selective anesthetic blockage of nerves or anatomical dissection is used. According to Sunderland, several factors cause difficulty in evaluating the motor function of the hand. The most important of these are anatomical variations or failure to evaluate the role of deceptive movements, given that these may imitate and cover up for loss of the original movements. If these factors are not properly valued, errors in making diagnoses and assessing the results will be inevitable.

Methods

Dissections were performed on 100 forearms of 50 adult cadavers belonging to our Anatomy Laboratory, among which 46 were male and four were female. Their ages ranged from 28 to 77 years; 27 were white and 23 were nonwhite. Forearms that had been deformed through injury, malformation or scarring were excluded.

The dissection was performed by means of a median incision along the entire forearm and the lower third of the upper arm. Two flaps that included the skin and subcutaneous tissue were pulled back toward the radial and ulnar sides, respectively. The same was done in relation to the fascia of the forearm, and thus the entire musculature was exposed (Fig. 1, Fig. 2).

Fig. 1

Median incision in the volar face along the entire forearm.

Fig. 2

Two flaps that included the skin and subcutaneous tissue were pulled back to expose the fascia and then the musculature of the forearm.

All of the muscles of the forearm were dissected. Their innervation and the presence of nerve communication between the forearm nerves were analyzed. All the anatomical variations encountered were recorded, annotated and photographed. A Keeler magnifying glass of 2.5× was used. A 1 cm segment of the anastomotic ramus was removed and taken to a surgical microscope. With the aid of microsurgical tweezers, the fascicles were separated, counted and photographed. In four limbs, we recorded only one fascicle; in eight, two; in nine, three; in four, four; in one, five; and in one, seven.

In addition to investigating Martin-Gruber nerve communication, the relationship between the Gantzer muscle and the anterior interosseous and median nerves and the anatomical variations of the forearm muscles were analyzed. This study was approved by the Research Ethics Committee.

Results

We recorded the presence of Martin-Gruber anastomosis in 27 of the 100 forearms dissected. Bilaterality was observed in three cadavers. In 13 cases, nerve communication occurred in the left forearm and in 14 cases in the right forearm. Regarding the topographic location, most of the nerve communications occurred in the proximal third of the forearm.

In the most frequently encountered form of communication, which was seen in nine forearms, the nerve fascicles originated from the anterior interosseous nerve, at a variety of sites, and headed distally at widely varying degrees of obliquity. They were located posteriorly to the ulnar artery and were supported on the deep flexors of the fingers, anteriorly to them. They headed toward the ulnar nerve and joined to it close to the junction between the proximal third and middle third of the forearm. In all of these cases, the fascicles originated distally to the elbow joint. We considered this to be type I (Fig. 3).

Fig. 3

Type I. This was the most frequent form of communication (seen in nine limbs). The nerve fascicles originated from the anterior interosseous nerve and headed toward the ulnar nerve.

Type II was recorded in two limbs. The nerve fascicles originated from the anterior interosseous nerve, just like in type I, with bifurcation of the fascicles, which communicated with the ulnar nerve at two different points (Fig. 4).

Fig. 4

In type II, which was recorded in two limbs, the nerve fascicles originated from the anterior interosseous nerve, with bifurcation, and communicated with the ulnar limb at two different points.

Type III, which was recorded in four anatomical specimens, differed from type I only in that it originated from the median nerve, close to the emergence of the anterior interosseous nerve. It had the same path as the preceding type, with variable obliquity. In three of these four cases, the origin of the nerve communication was proximal to the elbow joint (Fig. 5).

Fig. 5

In type III, which was recorded in four limbs, the fascicles originated from the median nerve, proximal to the emergence of the anterior interosseous nerve, and headed toward the ulnar nerve.

In type IV, which was recorded in five forearms, the communication occurred between the nerve fascicles that originated from the median and ulnar nerve branches and headed toward the deep flexor muscle of the fingers. These fascicles formed a loop with distal convexity. The fascicles of this communication headed toward the muscle mass of the deep flexors of the fingers (Fig. 6).

Fig. 6

In group IV, which was recorded in five limbs, communication occurred between nerve fascicles that were heading for the deep flexor muscle of the fingers, and the fascicles formed a loop with distal convexity.

Type V, which was also observed in five forearms, differed from the preceding type in that the communication between the fascicles coming from the median and ulnar nerves occurred inside the muscle mass of the deep flexors of the fingers. This was called intramuscular Martin-Gruber communication (Fig. 7).

Fig. 7

In type V, which was recorded in five limbs, nerve communication occurred inside the muscle mass of the deep flexors of the fingers (intramuscular anastomosis).

In type VI, which was recorded in two forearms, the fascicles originated from the branch of the median nerve toward the superficial flexor muscle and headed obliquely and distally to communicate with the ulnar nerve (Fig. 8).

Fig. 8

In type VI, which was recorded in two limbs, fascicles coming from the branch of the median nerve to the superficial flexor muscle headed toward communication with the ulnar nerve.

Segments of the anastomotic ramus of length 0.5–1.0 cm were removed. The number of fascicles was counted under a surgical microscope, at magnifications ranging from 10 to 16×. With the aid of microsurgical tweezers, the fascicles of the anastomotic ramus were separated. In four limbs, we recorded only one fascicle; in eight, two; in nine, three; in four, four; in one, five; and in one, seven (Fig. 9).

Fig. 9

Distribution of the number of fascicles in the 27 records of Martin-Gruber anastomosis. In four limbs, we only recorded one fascicle; in eight, two; in nine, three; in four, four; in one, five; and in one, seven.

Discussion

The incidence of nerve communication between the median and ulnar nerves in the forearm is variable and this becomes more marked when the results obtained through electroneuromyographic studies are compared with those from anatomical dissections. Table 1 shows the authors, the nature of the study, the number of cases and the percentage of nerve communication.

Table 1

Summary of the literature involving the incidence of anastomosis between the median and ulnar nerves (Martin-Gruber anastomosis), according to author.

Year	Authors	Study	Cases studied	Cases that presented Martin-Gruber anastomosis (%)
1870	Gruber4	Anatomical	250	15.2
1893	Thomson9	Anatomical	406	15.5
1931	Hirasawa10	Anatomical	No data	10.5
1976	Kimura et al.11	Electrophysiological	328	17
1981	Srinivasan and Rhodes12	Anatomical, genetic	200	15
1992	Uchida and Sugioka13	Electrophysiological	47 (cubital syndrome)	17
			87 (normal)	16
1992	Amoiridis26	Electrophysiological	100	32
1993	Nakashima14	Anatomical	108	21.3
1999	Shu et al.16	Anatomical, histological	72	23.6
2002	Rodriguez-Niedenfuhr et al.17	Anatomical	140	13.6
			236	13.1
2005	Lee et al.18	Anatomical, electrophysiological	102 (anatomical)	39.2
			224 (electrophysiological)	23.2
2015	Our results	Anatomical	100	27

To the best of our knowledge, no anatomical studies regarding the innervation of the opposing muscle of the thumb through the ulnar nerve have been conducted. However, through electromyographic studies, it was recorded by Sala in 32% of the cases, by Forrest in 20%, by Harness and Sequeles in 77%, by Highet in 20% and by Rosen in 16.7%. In comparing the studies of Sala and Forrest, in which exactly the same technique was used yet completely different results were achieved (Sala 77% and Forrest 20%), it can be seen that in the opinion of Forrest, this occurred because the authors placed electrodes at different depths and because Sala placed electrodes in contact with the muscle fibers of the short flexor muscle, thereby obtaining a higher percentage. This difference demonstrates that studies conducted through electrical stimulation are not as precise as those conducted through anatomical dissection.

Thompson was the first to describe a classification for Martin-Gruber anastomosis. Hirasawa was the first to report intramuscular Martin-Gruber anastomosis. Kimura et al., Srinivasan and Rhodes, Ucida and Sugioka, Nakashima, Oh, Shu et al., Rodriguez-Niedenfuhr et al. and Lee et al. also described classifications for this nerve anastomosis (Table 2). These classifications show many more differences than similarities between each other, which makes comparisons difficult. Thompson reported that in two of the 406 specimens dissected, the anastomotic ramus originated from the median nerve close to the ulnar nerve, in the distal third of the forearm. This anatomical format has so far not been recorded in the literature. It is probably an extremely rare anatomical variation, or a variation of group II of our findings, in which union with the ulnar nerve would occur in the distal third of the forearm.

Table 2

Summary of the literature involving different classifications for anastomosis between the median and ulnar nerves.

Year	Authors	Classification proposed; anastomosis between the median and ulnar nerves
1893	Thomson9	Class I: anterior interosseous nerve and ulnar nerve
Class II: median nerve and ulnar nerve
Class III: muscle branch to deep flexor muscle of the fingers
1931	Hirasawa10	Oblique anastomosis: anterior interosseous nerve and ulnar nerve
median nerve and ulnar nerve
Loop anastomosis: muscle branch to deep flexor muscle of the fingers
Combined anastomosis: combinations between others
1976	Kimura et al.11	Type I: median nerve and ulnar nerve innervating the hypothenar muscle
Type II: median nerve and ulnar nerve innervating the deep flexor muscle of the fingers
Type III: median nerve and ulnar nerve innervating the thenar muscle
1981	Srinivasan and Rhodes12	Types I, II, VI: anterior interosseous nerve and ulnar nerve or other
Type III: median nerve and ulnar nerve
Types IV, V: combinations of others
1992	Uchida and Sugioka13	Type I: median nerve and ulnar nerve innervating the hypothenar muscle
Type II: median nerve and ulnar nerve innervating the thenar muscle
Type III: median nerve and ulnar nerve innervating the deep flexor muscle of the fingers
1993	Nakashima14	Type Ia: anterior interosseous nerve and ulnar nerve
Type Ib: median nerve and ulnar nerve
Type III: combination of types Ia, Ib and II
1995	Oh et al.15	Type I: median nerve and ulnar nerve innervating the hypothenar muscle
Type II: median nerve and ulnar nerve innervating the deep flexor muscle of the fingers
Type III: median nerve and ulnar nerve innervating the thenar muscle
1999	Shu et al.16	Type I: anterior interosseous nerve and ulnar nerve
Type II: median nerve and ulnar nerve
Type III: muscle branch to deep flexor muscle of the fingers
Type IV: anterior interosseous nerve and ulnar nerve, muscle branch to deep flexor muscle of the fingers originating from the connection
Type V: two anastomotic branches
2002	Rodriguez-Niedenfuhr et al.17	Pattern I: one anastomotic ramus
Pattern II: two anastomotic rami
Type A: anastomotic ramus originating from a branch of the median nerve to the nerve of the superficial flexor muscle of the forearm
Type B: anastomotic ramus originating from the median nerve
Type C: anastomotic ramus originating from the anterior interosseous nerve
2005	Lee et al.18	Type I: anterior interosseous nerve and ulnar nerve
Type II: median nerve and ulnar nerve
Type III: muscle branch to deep flexor muscle of the fingers
Type IV: two anastomotic rami from the ulnar nerve or anterior interosseous nerve and ulnar nerve
2015	Our classification	Type I: anterior interosseous nerve and ulnar nerve
Type II: anterior interosseous nerve and ulnar nerve (double anastomosis)
Type III: median nerve and ulnar nerve
Type IV: loop between anterior interosseous nerve and ulnar nerve with branches to deep flexor muscle of the fingers
Type V: intramuscular anastomosis
Type VI: branch from the median nerve to the superficial flexor muscle and ulnar nerve

The classification drawn up by Srinivasan and Rhodes is the one that comes closest to our observations. We disagree with it only in relation to two points. These authors considered type I to be communications that originated in the proximal portion of the anterior interosseous nerve and group II to be those originating in the distal portion of this nerve. We observed that nerve communication could originate at different levels of the anterior interosseous nerve, but we did not see any reason to consider them to be different groups, since they originated in the same nerve and had the same format. The other point of disagreement is that these authors did not describe any occurrence of intramuscular anastomosis.

It is still not possible to identify the origin and destination of each nerve fascicle, either by electrophysiological or by anatomical means. Anatomical studies are based on the apparent path the fascicles are following, from one nerve to another. In agreement with the majority of authors,4, 5, 9, 10, 12, 14, 15, 16, 17 we can affirm that we did not observe any passage of motor fascicles from the ulnar nerve to the median nerve, in any of the limbs. We recorded that the communications were directed from the median nerve to the ulnar. However, nerves are composed of motor fascicles (efferent) and sensory fascicles (afferent). It is therefore possible that sensory fascicles (afferent) are directed from the ulnar to the median nerve. In group III, it is likely that in communications originating in the median nerve (which is a mixed nerve), there are both efferent (motor) and afferent (sensory) fascicles. In the other limbs, communication was by means of a motor nerve: anterior interosseous in types I, II, IV or V; or by means of a branch of the median nerve to the superficial flexor muscle in type VI. In this situation, it is unlikely that sensory fascicles will exist.

Shafic et al. reported the case of a patient who presented all of the symptoms of carpal tunnel syndrome, but had negative results from provocative tests such as Tinel's sign and Phalen's test. This patient presented evident signs of ulnar nerve compression in the elbow. Nerve stimulation in the epitrochlear-olecranon groove triggered the typical symptoms of carpal tunnel syndrome, which demonstrated that transfer of sensory nerve fascicles (afferent) from the ulnar nerve to the median nerve existed.

Marinacci described the case of a patient who suffered traumatic sectioning of the median nerve in the proximal forearm. An electromyographic evaluation showed denervation of the long flexor muscles of the thumb and the deep flexor of the index and middle fingers, which are usually innervated by the median nerve. However, the hand did not present motor alterations seen either clinically or electromyographically. From applying electromyographic stimuli and from the velocity of motor conduction of the median nerves in the wrist and elbow, it was concluded that this alteration occurred as a consequence of passage of nerve fibers from the ulnar nerve to the median nerve in the forearm, distally to the innervation of the long flexor of the thumb and the deep flexor of the index and middle fingers. The author named this type of innervation of all of the intrinsic muscles by the ulnar nerve “total ulnar hand”. Among our results, we recorded one limb with five and another with seven fascicles, but we cannot affirm that these fascicles would be sufficient to innervate all of the intrinsic muscles of the hand.

Using electrophysiological methods, Cliffton described a case in which all of the intrinsic muscles of the hand were found to be innervated and sensitivity was preserved, despite complete rupture of the ulnar nerve. This author also described another case in which there was complete rupture of the median nerve and all the intrinsic muscles were found to be functioning, despite the absence of sensitivity in the area of innervation of the median nerve. In the first situation, all the intrinsic muscles received innervation from the median nerve and in the second, from the ulnar nerve. These variations were attributed to Martin-Gruber anastomosis. Based on our studies, it is more likely that these variations occurred because of Cannieu-Riché anastomosis (communication from motor branches of the median nerve to the muscles of the thenar region and the deep branch of the ulnar nerve, in the palm of the hand), because this communication is three times more frequent than Martin-Gruber anastomosis.

Some reports in the literature have differentiated three types of Martin-Gruber communication. In types I and II, anomalous innervation of the abductor muscles of the little finger and the first dorsal interosseous muscle, respectively, are observed. In type III, there is anomalous innervation of the short abductor of the thumb. In types I and II, nerve fascicles head from the median to the ulnar nerve, while in type III they head from the ulnar to the median. Our view is that in these cases, the communication may have occurred either in the forearm or in the palm of the hand. It is more likely that this was Cannieu-Riché anastomosis, because this is three times more frequent than Martin-Gruber anastomosis.

Brandsma et al. described two clinical cases of patients with complete rupture of the ulnar nerve in the elbow and injury to the median nerve in the wrist, secondary to neuropathy due to leprosy, which maintained good functioning of the first dorsal interosseous muscles and the short flexor of the thumb. They attributed these findings to the presence of Martin-Gruber anastomosis. Our results contribute support to this clinical observation that this occurrence could only have been due to Martin-Gruber anastomosis.

Rosen and Lundborg examined 96 upper limbs of 48 patients. They found variations in the intrinsic muscles in 21% of the limbs: 75% in the hand and 25% in the forearm. They tested the electrical activity of the opposing muscles of the thumb and abductor of the little finger through stimulation of the ulnar and median nerves in the wrist and elbow. They concluded that in 11 patients there was transfer of nerve fibers from the ulnar to the median nerve. They did not record any nerve communication from the median to the ulnar nerve. These results are surprising, because they completely disagree with those in the literature, considering that they show ulnar–median anastomosis as being more frequent than median–ulnar.

Amoiridis reported being surprised about the results published by Golovchinsky, who found that nerve communication from the ulnar nerve to the median nerve was present in 16.7%. This author stated that Golovchinsky had not taken into consideration some basic principles of electrophysiological studies and that this was why the percentage of this type of nerve communication was so high. In addition, this author stated that Golovchinsky had not taken into consideration that a response from the elbow to the wrist diminished by up to 15% could be found even in normal individuals.

Komar et al. reported the case of a patient with complete rupture of the median nerve above the elbow, with a condition of incomplete paralysis of the median nerve in the hand and loss of sensitivity in the area of sensory innervation by the median nerve. They reported that their electromyographic evaluation suggested that nerve fiber transfer from the ulnar nerve to the median nerve in the forearm was present. Amoiridis stated that this case described by Komar was not one of nerve communication in the forearm, since stimulation in the wrist did not show a response from the muscles in the thenar region. We agree with Amoiridis, since it is likely that transfer of fascicles from the ulnar to the median nerve occurred distally to the wrist (Cannieu-Riché anastomosis).

Streib reported the case of a 77-year-old patient with a complaint of weakness in the hands. Electrical stimulation demonstrated that the response of the musculature of the thenar region had an amplitude greater than 50% when the median nerve was stimulated in the wrist, in relation to the elbow. The opposite occurred in relation to the ulnar nerve, i.e. the amplitude was 50% greater in the elbow. There is no doubt that in this case, nerve communication occurred in the forearm (Martin-Gruber anastomosis).

Dogan et al. described the case of a patient with a condition of compressive neuropathy of the ulnar nerve who presented altered sensitivity in the ring finger and little finger, but with atrophy of the short abductor muscle of the thumb. They attributed this finding to an anomaly of innervation of the short abductor.

The main classical treatises describe Martin-Gruber anastomosis as nerve communication in the forearm between the median and ulnar nerves, but do not mention either the origin or the destination of the nerve fibers.

Conclusion

Knowledge of the anatomical variations relating to innervation of the hand is particularly important with regard to performing physical examinations, making diagnoses, determining the prognosis and implementing surgical treatment. If these variations are not given their proper value, errors and consequences will be inevitable.

Nerve communications between the median and ulnar nerves may cause alterations to clinical symptoms, especially in patients with carpal tunnel syndrome, since these variations may exacerbate or attenuate the symptoms and cause motor and sensory alterations that differ from the usual pattern. Clinically, variations in motor innervation are very important, in that even with complete rupture of the median or ulnar nerve, some of these muscles may not necessarily become paralyzed. Such situations could lead to the wrong conclusion that the nerve had not become completely severed. It is essential to differentiate between partial and total rupture. However, correct identification of these variables is not always easy, since rigorous clinical and electromyographic examination is needed. Alterations recorded in the electroneuromyographic examination without clinical proof are insufficient for indicating a surgical procedure. Careful dissection during the surgical procedure may prevent injury to an anastomotic ramus.

Conflicts of interest

The authors declare no conflicts of interest.