Extensile operative exposure of the posterior aspect of the humeral diaphysis is limited by the course and mobility of the radial nerve. We are aware of only a few studies that have attempted to define both the precise location of the radial nerve along the posterior aspect of the humerus, with reference to osseous landmarks, and the branches of the radial nerve to the triceps in this location1,7,12,14,21,23.
An anatomical study was performed not only to investigate the anatomical relationship of the radial nerve to the posterior aspect of the humerus but also to define the local branching of the nerve, which influences its mobility. On the basis of the anatomical findings, operative approaches to the posterior aspect of the humeral diaphysis were studied, with particular reference to the limitation imposed by the radial nerve on exposure of the humerus.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†The Hospital for Special Surgery, 535 East 70th Street, New York, N.Y. 10021.
Cadaveric Series
Ten specimens were obtained at autopsy within seventy-two hours after death; seven of the specimens were from women and three were from men. Only right upper extremities were used. The posterior aspect of the arm was dissected with the torso supine and the shoulder in 90 degrees of flexion, full internal rotation, and neutral abduction. The elbow was flexed, to allow the forearm to drape across the body, duplicating a commonly used position for the posterior approach to the arm. A longitudinal incision was made in the midline of the posterior aspect from eight centimeters distal to the acromion to the olecranon fossa. The deep fascia of the arm was incised in line with the skin incision. The common tendon of the lateral and long heads of the triceps was identified and split longitudinally, permitting the long head of the triceps to be retracted medially and the lateral head to be retracted laterally. Measurements were made of the course of the radial nerve as it crossed the posterior aspect of the humerus, with the medial and lateral epicondyles used as distal reference points. The length of the humerus from the medial epicondyle to the medial aspect of the anatomical neck was also documented for each specimen.
The same specimens were used to examine the extent of the posterior aspect of the humeral diaphysis that could be visualized through three operative approaches. The sequence of the approaches was such that one approach did not compromise the others. The most proximal extent of the exposed humerus was marked, and measurements were made from this point to the part of the lateral epicondyle that was most palpable subcutaneously.
The first approach was the posterior triceps-splitting approach12. The medial head of the triceps was split through the previously created posterior skin incision down to the periosteum of the humerus, and the dissection was continued subperiosteally around the humerus (Fig. 1-A). Measurements were then made from the most prominent point of the lateral epicondyle to the most proximal aspect of the exposed diaphysis up to the point at which the radial nerve crossed the operative field.
The second operative approach was a modification of the first wherein the lateral head of the triceps was elevated laterally. The radial nerve was then elevated in a superior direction limited by the tethering effect of the branch to the medial head of the triceps (Fig. 1-B). The most proximal extent of the exposure of the humeral diaphysis was measured from the lateral epicondyle, as previously described.
With the third approach, the triceps muscle was retracted medially to expose the lower lateral brachial cutaneous nerve that branches off of the radial nerve on the posterior aspect of the lateral intermuscular septum (Fig. 2-A). This branch was traced proximally to identify the main trunk of the radial nerve proximal to the point at which it pierced the intermuscular septum. The intermuscular septum was divided distally for three centimeters over the radial nerve to permit operative mobilization of the nerve (Fig. 2-B). The medial and lateral heads of the triceps muscle were then elevated subperiosteally and retracted medially. The radial nerve was similarly retracted medially with the elevated triceps muscle to allow exposure of the humeral diaphysis. This exposure was extended to the axillary nerve (Fig. 2-C).
After the three exposures of the humerus had been made, the lateral head of the triceps was divided at the point at which the radial nerve crossed the posterior aspect of the humerus. The course and branching of the radial nerve from its entrance into the posterior compartment, below the teres major tendon, to its exit through the lateral intermuscular septum was then examined. The number and location of any branches arising from the nerve were recorded.
Clinical Series
The cases of seven patients who had had a modified posterior approach as treatment for a non-union (four patients) or an acute fracture of the humerus (three patients) were reviewed. All of the fractures had occurred in the middle third of the humerus, with some extending into the distal third. Two of the acute fractures and three of the non-unions were short oblique and one each was transverse. Previous operative procedures aimed at obtaining union had been performed for two of the non-unions. One patient who had an acute fracture had had a radial nerve palsy at presentation.
Radiographic measurements made for these seven patients included the length of the humerus from the medial epicondyle to the medial aspect of the anatomical neck, the proximal extent of the fracture measured from the lateral epicondyle to the most proximal point of the fracture, and the proximal extent of plate fixation measured from the lateral epicondyle to the proximal aspect of the humeral plate. All measurements were rounded to the nearest half centimeter.
In five patients, the modified posterior approach was used because the location of the acute fracture or non-union was proximal to the point at which the standard posterior triceps-splitting approach allows sufficient osseous exposure to apply a plate. In one patient who had a non-union two years after an injury, the initial operative dissection was performed through a standard triceps-splitting approach. Abundant fibrosis at the site of the non-union made dissection difficult, and a modified posterior approach was then used. The radial nerve was identified distally, and it was found to extend into the fibrosis at the site of the non-union as it was traced proximally. One patient had had a preoperative radial nerve palsy following attempted reduction of a humeral fracture in the emergency room. The modified posterior approach was used to identify the radial nerve and to allow for internal fixation of the humerus. In this patient, the nerve was noted to be tethered on the lateral aspect of the distal fragment.
Internal fixation was applied beneath the radial nerve after exposure with the modified posterior approach in all of the patients. When the triceps muscle was allowed to resume its normal position, the dissected part of the radial nerve resumed its lateral position and was not in contact with the plate. Instead, the medial and lateral heads of the triceps lay over the plate with the undissected part of the radial nerve.
Cadaveric Specimens
Anatomy of the Nerve
The average length (and standard deviation) of the humerus from the medial epicondyle to the medial aspect of the anatomical neck was 28.0 ± 1.9 centimeters. The radial nerve ran along the posterior aspect of the humerus from 20.7 ± 1.2 centimeters proximal to the medial epicondyle (74 per cent of the length of the humerus) to 14.2 ± 0.6 centimeters proximal to the lateral epicondyle (51 per cent of the length of the humerus) (Fig. 3). Throughout this course along the posterior aspect of the humerus, the radial nerve was adjacent to the bone without an intervening layer of muscle for an average span of 6.5 centimeters. The nerve gave off multiple branches to the lateral head of the triceps but none to the medial head of the triceps. At the lateral aspect of the humerus, the nerve trifurcated into a branch to the medial head of the triceps, the lower lateral brachial cutaneous nerve, and the continuation of the radial nerve into the forearm. After the trifurcation, the radial nerve pierced the intermuscular septum an average of 10.2 ± 0.4 centimeters proximal to the lateral epicondyle (the distal 36 per cent of the humerus).
Exposure of the Humerus
With the standard posterior triceps-splitting approach without mobilization of the radial nerve, the proximal dissection extended to a point where the nerve crossed the posterior aspect of the humerus at an average of 15.4 ± 0.8 centimeters proximal to the lateral epicondyle (the distal 55 per cent of the humerus). With mobilization of the radial nerve proximally and elevation of the lateral head of the triceps, the exposure of the humerus was increased to an average of 21.4 ± 1.2 centimeters from the lateral epicondyle (the distal 76 per cent of the humerus). With the modified posterior approach of subperiosteal reflection of the medial and lateral heads of the triceps medially, 26.2 ± 0.4 centimeters of the distal aspect of the humerus was exposed (the distal 94 per cent of the humerus) (Fig. 4).
Clinical Series
The average duration of the postoperative follow-up was twenty-nine months (range, twenty-one to thirty-eight months). No patient had any evidence of dysfunction of the radial nerve or clinical evidence of weakness of the triceps at the time of follow-up. The patient who had had a radial nerve palsy regained complete sensory and motor function by six months postoperatively.
Union was achieved in six of the seven patients. The one failure occurred in a patient who had had a previous attempt at internal fixation with a plate. Although the operative exposure was adequate, the non-union was thought to be the result of unstable internal fixation. Union was achieved later, after the application of an Ilizarov external fixator frame.
The average length of the humerus from the medial epicondyle to the medial aspect of the anatomical neck was 27.6 ± 1.1 centimeters (range, 24.5 to 31.5 centimeters). The average proximal extent of the fracture was 15.6 ± 0.9 centimeters (range, 13.0 to 19.0 centimeters) from the lateral epicondyle or 57 per cent of the length of the humerus. The average proximal extent of the plate fixation was 22.1 ± 1.0 centimeters (range, 18.5 to 27.0 centimeters) from the lateral epicondyle or 80 per cent of the length of the humerus.
Operative treatment of humeral fractures is effective for patients who have multiple traumatic injuries, vascular injuries, or a non-union2,4-6,9,17. An inability to obtain a closed reduction in a patient who has a radial nerve palsy due to entrapment of the nerve at the site of the fracture6 or a completely transected nerve associated with an open humeral fracture6,13 make visualization of the nerve desirable. Knowledge of the location of the radial nerve with reference to the posterior aspect of the humerus and its influence on the potential proximal extent of the operative approach is of major importance.
A review of the literature demonstrated little information on the anatomy of the radial nerve, although its injury in association with humeral fractures has been well described1,3,11,15,16,18-20,22. Whitson reported on the proximity of the nerve to the bone but not on the anatomy of the nerve or its relationship to the longitudinal axis of the humerus. His study was directed at explaining the cause of neural injury in association with humeral fractures and not at detailing operative approaches to the posterior aspect of the humerus. In some of his dissections, the radial nerve ran within a thin layer of muscle. In our dissections, we found no intervening layer of muscle between the nerve and the bone.
According to Henry, in its course along the posterior aspect of the humerus, the radial nerve has one large branch to the lateral head of the triceps and one large branch to the medial head of the triceps running within the posteromedial aspect of the muscle belly. Henry's illustrations also suggested that the radial nerve crosses the humerus at the level of its proximal third. In other texts7,8, diagrams of the radial nerve have shown multiple branches to the medial and lateral heads of the triceps. Sunderland found that one-third of the medial heads of the triceps and one-half of the lateral heads of the triceps that he dissected were supplied by branches of the radial nerve that arose as the nerve traversed along the posterior aspect of the humerus. This was in contrast to the findings of Linell, in 1921, that the lateral head of the triceps was supplied by numerous subdivisions of one branch and that the medial head was supplied by one branch that ran for seven centimeters extramuscularly. The location of this course relative to the long axis of the humerus was not described.
Branches of the radial nerve to the medial head of the triceps along the posterior aspect of the humerus would make mobilization of the radial nerve in this region difficult. Our dissections demonstrated no such branches until the nerve crossed the lateral aspect of the humerus. It is because of this anatomy that the nerve can be elevated proximally.
The modified posterior approach, in addition to providing the greatest visualization of the posterior aspect of the humerus, has an added benefit of reflecting the medial head of the triceps en masse and maintaining continuity of the fibers. If the triceps-splitting approach is attempted but inadequate exposure is obtained, the modified posterior approach can be performed through the same incision. The splitting of the medial head of the triceps is abandoned, the triceps is retracted medially, and the lower lateral brachial cutaneous nerve is identified branching off of the radial nerve at the lateral aspect of the triceps. The modified posterior approach is then performed, as previously described. Internal fixation of the humerus is obtained with a compression plate deep to the radial nerve. The medial and lateral heads of the triceps are allowed to resume their normal position with excellent coverage of the plate.
Knowledge of the specific location of the radial nerve relative to the lateral epicondyle can assist in the decision regarding the operative approach to the posterior aspect of the humerus. Although technically difficult, the modified posterior approach described here allows for increased visualization of the posterior aspect of the humerus, with identification and protection of the radial nerve and without division or denervation of the medial head of the triceps. Since almost the entire posterior aspect of the humerus can be exposed, this approach allows for greater flexibility when internal fixation is used in the humerus.