Twenty upper extremities from twenty fresh-frozen human cadavera (mean age at the time of death, seventy-five years; range, forty-eight to ninety-eight years) were obtained by disarticulation at the scapulothoracic joint. The shoulder girdle with the axillary sheath and the cords of the brachial plexus were preserved in order to approximate the resting length of and tension on the ulnar nerve with the shoulder in the neutral position. The proximal end of the ulnar nerve was not constrained. Physical examination of the extremities indicated no evidence of subluxation of the ulnar nerve and a normal range of motion of the elbow for each specimen. The specimens were stored in plastic bags at -20 degrees Celsius and then were thawed to room temperature (24 degrees Celsius) for imaging or pressure measurement. The first ten consecutive specimens (mean age at the time of death, seventy-two years; range, forty-eight to eighty-nine years) were used for evaluation of the cubital tunnel with magnetic resonance imaging. The second ten specimens (mean age at the time of death, seventy-eight years; range, sixty-four to ninety-eight years) were used for evaluation of intraneural and extraneural pressures in the cubital tunnel.
Magnetic Resonance Imaging
The specimens were thawed and placed in an acrylic fixture that allowed consistent positioning of the elbow at 0, 45, 90, or 135 degrees of flexion during imaging. The wrist was maintained at 0 degrees of pronation-supination and 0 degrees of flexion. The fixture was positioned in a 1.5-tesla Magnetom Vision scanner (Siemens Medical Systems, Iselin, New Jersey). A circular polarized small flex coil (Siemens) was placed around the elbow for signal enhancement. Imaging was performed with use of a standard T1 protocol, with an echo time of fifteen milliseconds and a repetition time of 500 milliseconds.
We considered three anatomical regions of the cubital tunnel, as described by Eversmann, and analyzed each region individually. The first region (A) was at the entrance of the tunnel just posterior to the medial epicondyle, the second region (B) was at the aponeurosis joining the two heads of the flexor carpi ulnaris, and the third region (C) was at the proximal muscle belly of the flexor carpi ulnaris (Fig. 1). Cross-sectional images were made at a defined level in each region, in a plane perpendicular to the local tunnel axis (Fig. 1). For region A, the section was located at the ulnar groove on the posterior aspect of the medial epicondyle, three millimeters proximal to the trochlea and perpendicular to the long axis of the humerus. For region B, the section was located three millimeters distal to the proximal surface of the trochlea through an axis bisecting the elbow joint. For region C, the section was located three millimeters proximal to the radial head, perpendicular to the long axis of the ulna.
Cross-sectional areas of the cubital tunnel and the ulnar nerve were determined by quantitative analysis of the magnetic resonance images. Image files were transferred to a computer (Apple Macintosh, Cupertino, California) and were analyzed with image-processing software (NIH Image; National Institutes of Health, National Technical Information Service, Springfield, Virginia). The outlines of the cubital tunnel and the ulnar nerve were traced interactively on each image (Fig. 2). The demarcation of specific structures was unambiguous because of the high contrast between the nerve, the perineural fat, and the fibro-osseous cubital tunnel. All measurements were performed by two of us (F. P. H., Jr., and R. R. B.), and the results were averaged. Measurements of the area of the cubital tunnel and the ulnar nerve were obtained from sections of the three regions of the cubital tunnel with the elbow in 0, 45, 90, and 135 degrees of flexion. To quantify the proportion of the tunnel that was occupied by the nerve, the ratio of the area of the nerve to that of the cubital tunnel (the nerve-tunnel area ratio) was calculated. In addition to the quantitative analysis, the shapes of the tunnel and the nerve were noted as a function of the region and of the angle of flexion of the elbow.
The effect of the angle of the elbow on the area of the tunnel, the area of the nerve, and the nerve-tunnel area ratio was examined with one-way analysis of variance (JMP software; SAS Institute, Cary, North Carolina). Measurements for each region were compared individually with use of a Tukey multiple-comparison test, with a significance level of 0.05.
Intraneural and Extraneural Pressures
For measurements of pressure, the specimens were placed in a fixture that allowed for simultaneous measurement of angulation of the elbow and interstitial pressure, similar to the one described by Wright et al. The limb was positioned in the horizontal plane, with the lateral surface of the forearm resting against the base of the fixture to allow access to the cubital tunnel region. The limb was fixed rigidly to the apparatus with two five-millimeter Steinmann pins placed across the humerus in an anteromedial-to-posterolateral direction in a way that avoided impingement on the ulnar nerve. The fixture allowed for an arc of motion of the elbow from 0 to 130 degrees with 10-degree gradations. The wrist was maintained at 0 degrees of pronation-supination and 0 degrees of flexion.
Pressure transducers were positioned under ultrasonographic visualization (Fig. 3). Ultrasonographic imaging was performed with one of two systems with use of high-resolution linear-array transducers (HDI 3000; Advanced Technologies Laboratories, Bothell, Washington, or 128 XP 10; Acuson, Mountain View, California). The medial epicondyle, the olecranon, and the arcuate ligament were visualized, and the ulnar nerve was identified within the cubital tunnel and at a level four centimeters proximal to it. An 18-gauge needle was inserted percutaneously and advanced until the tip was at the desired location for measurement of pressure. A flexible slit catheter (Stryker, Kalamazoo, Michigan) was inserted through the needle, and the needle was withdrawn. The catheter was anchored to the skin of the brachium with adhesive tape. Pressures were measured with a pressure transducer (Spectramed; Gulton Stratham, Costa Mesa, California) and were recorded on a strip-chart recorder (Gould Instruments, St. Sauveur, Ballainvilliers, France). The catheter and transducer were flushed with the minimum volume of saline solution required to obtain continuity between the tip of the catheter and the interstitium. Pressures in each intact arm were recorded at four locations—in the ulnar nerve within the cubital tunnel (region B, Fig. 1), in the ulnar nerve four centimeters proximal to the cubital tunnel, adjacent to the ulnar nerve within the cubital tunnel, and adjacent to the ulnar nerve four centimeters proximal to the cubital tunnel—with the catheter positioned at approximately the same elevation each time. Measurements were recorded at each location, at 10-degree increments from 0 to 130 degrees of elbow flexion. After each series of measurements, ultrasonography was used to confirm that the catheter had not become dislodged. The measurements were performed three times without removal of the catheter, and the mean of the three trials was recorded, as in previous studies4,19.
The ulnar nerve then was operatively released by incision of the subcutaneous tissues, the aponeurotic arch of the cubital tunnel, and the fibrous arcades between the two heads of the flexor carpi ulnaris, as described by Ferlic. Measurements of intraneural pressure were repeated at the level of the cubital tunnel and four centimeters proximal to it as the elbow was flexed from 0 to 130 degrees.
Three statistical analyses were performed on the mean values for ten specimens, with use of the two-tailed, paired Student t test and a significance level of 0.05. First, at each level and for each angle of flexion, the extraneural pressure and the intraneural pressure after release were compared with the intraneural pressure in the intact specimen. Second, at each level, the intraneural and extraneural pressures that were measured with the elbow at 10 to 130 degrees of flexion were compared with those measured with the elbow at 0 degrees. Third, at each level, the intraneural and extraneural pressures that were measured with the elbow at 0 to 130 degrees were compared with those measured with the elbow at the position of lowest mean pressure.
Magnetic Resonance Imaging
Satisfactory images were obtained for all ten specimens. On each image, the cubital tunnel was identified at the posteromedial aspect of the elbow by noting the high signal intensity of the contained fat (Fig. 2). The boundaries of the tunnel were well demarcated by the intermediate signal intensity of the surrounding muscle and retinaculum. The ulnar nerve was identifiable by a homogeneous intermediate-level signal with good contrast relative to the surrounding fat.
The cross-sectional shape of the cubital tunnel depended on the degree of flexion of the elbow and varied slightly between regions. In regions A and B, the tunnel was approximately triangular, with the base formed by the lateral and inferior surfaces of the medial epicondyle, with the elbow flexed 0 and 45 degrees, respectively (Fig. 4). In region A, the olecranon and the medial edge of the triceps formed the posterolateral border and a layer of fascia formed the posteromedial border. In region B, the aponeurosis joining the two heads of the flexor carpi ulnaris formed the posteromedial border and the elbow joint formed the posterolateral border. With increasing flexion of the elbow, to 90 and 135 degrees, the shape of the tunnel became elliptical, with approximation of the medial and lateral borders. In region C, with the elbow at 0 degrees, the tunnel was circular and was surrounded by soft tissue on all sides. The tunnel appeared triangular with the elbow flexed 45 degrees and elliptical at 90 and 135 degrees. In summary, in each of the three regions the cubital tunnel appeared triangular or circular in cross section with the elbow flexed 0 and 45 degrees, but it became elliptical at 90 and 135 degrees.
The position of the ulnar nerve within the tunnel did not change with flexion of the elbow, although the width of the nerve appeared to decrease with increasing flexion. At all positions of the elbow, the ulnar nerve occupied the anterior portion of the tunnel in regions A and B, directly apposed to the medial epicondyle anteriorly and surrounded by fat elsewhere, and the nerve occupied the central portion of the tunnel in region C, where it was surrounded by fat on all sides. There was no evidence that the walls of the tunnel collapsed and impinged directly on the nerve at any position of the elbow.
The cross-sectional areas of the cubital tunnel and the ulnar nerve decreased with increasing flexion of the elbow (Table I and Fig. 5). In region A, the mean area of the tunnel decreased by 30 per cent as the elbow flexed from 0 to 135 degrees and the mean area of the nerve decreased by 33 per cent. With the elbow flexed 90 and 135 degrees, the mean area of the tunnel was significantly smaller than it was at 0 degrees (p = 0.04 and p = 0.02, respectively) and the mean area of the nerve was significantly smaller than it was at 0 degrees (p < 0.001 for each comparison) and at 45 degrees (p = 0.04 and p = 0.005, respectively). In region B, the mean area of the tunnel decreased by 39 per cent as the elbow flexed from 0 to 135 degrees and the mean area of the nerve decreased by 50 per cent. With the elbow flexed 135 degrees, the mean area of the tunnel was significantly smaller than it was at 0, 45, and 90 degrees (p < 0.001, p = 0.001, and p = 0.03, respectively); the mean area of the nerve also was significantly smaller than it was at those positions (p < 0.001, p < 0.001, and p = 0.05, respectively). In region C, the mean area of the tunnel at 135 degrees was 41 per cent smaller than it was at 0 degrees; however, with the numbers available, this difference was not found to be significant (p = 0.10). The interspecimen variability in the area of the tunnel was greatest in region C, perhaps because of the divergent shape of the tunnel in this region. When the elbow flexed from 0 to 135 degrees, the mean area of the nerve in this region did decrease significantly (p = 0.04), by 34 per cent. The nerve-tunnel area ratios, as measured in all three regions and at all angles of flexion, ranged from a mean (and standard deviation) of 0.36 ± 0.11 to 0.50 ± 0.13. However, with the numbers available, we could detect no significant effect of the angle of flexion, indicating that the areas of the nerve and tunnel decreased by approximately equal proportions with increasing flexion.
Intraneural and Extraneural Pressures
The relationship between intraneural and extraneural pressures within the cubital tunnel of the ten specimens depended on the position of the elbow (Table II and Fig. 6). The mean intraneural pressure was significantly higher than the mean extraneural pressure with the elbow at 90, 100, 110, and 130 degrees of flexion (p < 0.05 for each comparison). The greatest difference between intraneural and extraneural pressure (41.0 ± 7.8 compared with 28.3 ± 14.6 millimeters of mercury [5.47 ± 1.04 compared with 3.77 ± 1.95 kilopascals]; p < 0.001) was seen at 130 degrees, the position at which these pressures were highest.
With the elbow at 30 to 70 degrees of flexion, the mean intraneural and extraneural pressures within the cubital tunnel were significantly lower than they were with the elbow in full extension (p < 0.05 for each comparison). The lowest mean intraneural pressure (2.0 ± 1.4 millimeters of mercury [0.27 ± 0.19 kilopascal]) was recorded with the elbow flexed 50 degrees, and the lowest mean extraneural pressure (2.5 ± 2.3 millimeters of mercury [0.33 ± 0.31 kilopascal]) was recorded with the elbow flexed 40 degrees. Compared with the lowest mean pressures, the intraneural pressure increased significantly with the elbow flexed 70 degrees or more, whereas the extraneural pressure increased significantly with the elbow flexed 100 degrees or more (p < 0.05 for each comparison). Compared with the mean values obtained with the elbow in full extension, the intraneural mean pressure increased significantly with the elbow flexed 100 degrees or more and the mean extraneural pressure increased significantly with the elbow flexed 110 degrees or more (p < 0.05 for each comparison).
The relationship between intraneural and extraneural pressure four centimeters proximal to the cubital tunnel also depended on the position of the elbow (Table III and Fig. 7). With the elbow at 120 and 130 degrees of flexion, the mean intraneural pressure was significantly greater than the mean extraneural pressure (p < 0.05 for each comparison). The greatest difference (22.2 ± 7.5 compared with 13.0 ± 6.4 millimeters of mercury [2.96 ± 1.00 compared with 1.73 ± 0.85 kilopascals]; p < 0.001) was seen with the elbow at 130 degrees, the position at which the pressures were the greatest. Compared with the values obtained with the elbow in full extension, the mean intraneural pressure increased significantly with the elbow flexed 110 degrees or more and the mean extraneural pressure increased significantly with the elbow flexed 120 degrees or more (p < 0.05 for each comparison).
After release of the aponeurosis of the cubital tunnel, the ulnar nerve subluxated anteriorly in five of the ten limbs and only the five nerves that did not subluxate were considered for additional analysis. With the numbers available, the mean intraneural pressures in the cubital tunnel after release were not found to be significantly different from those before release (p > 0.05 for each comparison). Compared with the value obtained with the elbow in full extension, the mean intraneural pressure did not decrease initially with flexion, as it had in the intact specimens; however, it did increase significantly when the elbow was flexed 90 degrees or more (p < 0.05 for each comparison), as it had in the intact specimens.
After operative release of the ulnar nerve, the mean intraneural pressures four centimeters proximal to the cubital tunnel were not found to be significantly different, with the numbers available, from those measured in the intact specimens (p > 0.05 for each comparison). The intraneural pressure increased significantly with the elbow flexed 90 degrees or more compared with that recorded with the elbow in full extension (p < 0.05 for each comparison).
Narrowing of the fibro-osseous cubital tunnel as the elbow flexes and traction-related deformation of the nerve as it curves over the medial epicondyle have been proposed as the most likely causes of compression of the ulnar nerve in this region. Apfelberg and Larson examined molds of acrylic that had been injected into the cubital tunnel with the elbow in extension and flexion and observed a change in the shape of the tunnel from round to elliptical and a mean reduction in the height from 5.8 to 3.2 millimeters. Those authors1 and others6,11,13,16,17 have attributed decreases in the area of the cubital tunnel to increasing tension of the aponeurotic roof. Bulging of the medial collateral ligament in the floor of the cubital tunnel has also been reported to decrease the size of the tunnel1,6,17. Measurements of interstitial pressure in the cubital tunnel have provided additional data indirectly indicating that extrinsic compression of the ulnar nerve may occur as flexion of the elbow increases. Macnicol recorded extraneural pressures at two sites in the cubital tunnel with the elbow at different positions. He found that, with maximum flexion of the elbow, pressure increased from 5.0 millimeters of mercury (0.67 kilopascals) to 40.0 and 47.0 millimeters of mercury (5.33 and 6.27 kilopascals) in the postcondylar groove and the region deep to the aponeurosis of the flexor carpi ulnaris muscle, respectively.
The mechanical effects of the position of the elbow on the ulnar nerve have not been determined conclusively. Although Apfelberg and Larson reported elongation of the ulnar nerve by 4.7 millimeters adjacent to the elbow with full flexion, they did not indicate the segment of the nerve that was studied. In a recent study, elongation of the ulnar nerve in association with flexion of the elbow was recorded proximal and distal to the retinaculum of the cubital tunnel15. An 18 per cent elongation of the nerve, predominantly in the proximal segment, was observed in association with flexion of the elbow to 135 degrees15. Pechan and Julis studied changes in intraneural pressure during flexion of the elbow in an attempt to determine the magnitude of traction-related deformation. They recorded a modest increase in intraneural pressure, from 7.2 to 11.1 millimeters of mercury (0.96 to 1.48 kilopascals), in association with flexion of the elbow to 90 degrees. The application of additional stretch on the ulnar nerve by dorsiflexion of the wrist, flexion of the elbow, and elevation of the shoulder resulted in a marked increase in intraneural pressure to 45.7 millimeters of mercury (6.09 kilopascals)13. In summary, previous studies1,6,11,13,15-17 have suggested that both external compression by the walls of the cubital tunnel and stretching of the ulnar nerve can occur progressively as flexion of the elbow increases, but the relative contributions of these two factors to changes in ulnar nerve pressure are not known.
Because of the promising findings in previous reports, we used magnetic resonance imaging of the elbow to obtain cross-sectional images of the cubital tunnel and the ulnar nerve in three regions. After developing and standardizing the protocol to obtain consistent images between specimens, we noted advantages of magnetic resonance imaging compared with other techniques. The ulnar nerve and the cubital tunnel could be simultaneously examined in situ. It was possible to select representative sections for examination of each region of the cubital tunnel and to identify reproducible levels and axes of section for procurement of comparable images at different positions of the elbow. Because the ulnar nerve and the cubital tunnel were studied non-invasively, the changes that were observed with flexion of the elbow are more likely to approximate those occurring in vivo.
Similarly, we were able to evaluate interstitial pressures in the cubital tunnel and the ulnar nerve in a minimally invasive manner with use of ultrasonographic imaging to visualize the insertion of the pressure catheter. Previous investigators have used dissection methods to place pressure-measuring devices in either the cubital tunnel or the ulnar nerve4,8,13,18. The resultant alterations in the anatomical structure may have affected the pressures that were recorded. We eliminated this possible source of error by using high-resolution ultrasonographic imaging, and this was demonstrated to be a useful tool for future studies of neural pressure.
We found that, within the cubital tunnel, the mean intraneural pressure was significantly greater than the mean extraneural pressure when the elbow was flexed 90 degrees or more (p < 0.05). With the elbow flexed 130 degrees, the mean intraneural pressure was 45 per cent greater than the mean extraneural pressure (Table II). Similarly, with the elbow flexed 120 degrees or more, the mean intraneural pressure four centimeters proximal to the cubital tunnel was significantly greater than the mean extraneural pressure (p < 0.01). Relative to the lowest values, the intraneural pressure increased significantly at smaller angles of flexion than did the extraneural pressure, both within the cubital tunnel and proximal to it. For example, the intraneural pressure within the cubital tunnel increased significantly with the elbow flexed 70 degrees (p < 0.05), whereas the extraneural pressure did not increase significantly until the elbow was flexed 100 degrees. These data indicate that the increase in intraneural pressure that occurs with flexion of the elbow is not due entirely to increased extraneural pressure.
The increases in interstitial pressure that were noted in association with flexion of the elbow corresponded closely with findings on magnetic resonance imaging indicating that consistent alterations in the cross-sectional area of the cubital tunnel and the ulnar nerve occur with the elbow flexed 90 degrees or more. The finding that perineural fat surrounded the ulnar nerve in all positions of flexion is analogous to the visualization of epidural fat in magnetic resonance imaging studies of the lumbar spine and may be considered an indicator of free space around the nerve9. These observations do not support the concept that direct, focal compression of the ulnar nerve occurs because of a disproportionate decrease in the area of the cubital tunnel compared with that of the ulnar nerve within the aponeurotic region of the tunnel in the normal elbow. Alternatively, the observation that the cross-sectional area of the ulnar nerve decreases and the interstitial pressure increases significantly without direct extrinsic compression suggests that traction-related deformation occurs within the ulnar nerve as the normal elbow flexes.
We acknowledge several limitations of this study. As our specimens had none of the pathological characteristics associated with entrapment of the ulnar nerve at the elbow, the results for these normal elbows may not apply directly to patients who have cubital tunnel syndrome. For example, if extraneural pressure is increased abnormally in an elbow with cubital tunnel syndrome compared with that in a normal elbow, release of the aponeurosis may decrease intraneural pressure; our data suggest that there would be no effect in a normal elbow. A second limitation is that we used forequarter amputation specimens and did not attempt to reproduce the in vivo traction on the proximal end of the ulnar nerve. Therefore, the experimental conditions probably resulted in decreased proximal traction compared with the in vivo state. However, even under these conditions, we observed changes in morphology and pressure that are consistent with a mechanism of increased traction secondary to flexion of the elbow. It is likely that increased tension on the ulnar nerve proximally would have further increased the traction-related deformation effects that we observed with flexion of the elbow.
The consistent observation that cross-sectional area decreased and interstitial pressure increased within the cubital tunnel and the ulnar nerve with flexion of the elbow may have clinical relevance. With the elbow flexed 90 and 135 degrees, the mean area of the cubital tunnel in the subaponeurotic region decreased (by 18 and 39 per cent) compared with the area at full extension, and the ulnar nerve underwent similar decreases (24 and 50 per cent). These data are supported by the finding that intraneural and extraneural pressures within the cubital tunnel are lowest at 50 and 40 degrees, respectively, and they indicate that the optimum position for immobilization of the elbow, with respect to neural pressure and cross-sectional area, is approximately 45 degrees of flexion. As there have been no reports of pressures measured with the elbow between 0 and 90 degrees of flexion, to our knowledge, this phenomenon may have been overlooked previously. In a recent magnetic resonance imaging study, the ulnar nerve was observed to buckle on itself when the elbow was in full extension and to straighten with progressive flexion12. Our data indicate that pressures increase slightly with the elbow in positions of extension where the nerve may be redundant and that they decrease with the elbow in the initial stages of flexion as the nerve elongates. Pressures increase as flexion increases to more than 90 degrees, a finding that corresponds well with the observed decrease in cross-sectional areas seen on magnetic resonance imaging.
If the findings in the current study of normal human cadaveric elbows are supported in studies of elbows with entrapment of the ulnar nerve, it may be concluded that operative procedures that decompress the ulnar nerve without either transposing it or decompressing it through a medial epicondylectomy do not effectively relieve the symptoms caused by neural traction. As neural traction appears to be a source of increased intraneural pressure under normal conditions, it appears reasonable that a goal of treatment of entrapment neuropathy is the elimination of all sources of nerve impingement, including those causing excessive neural traction. These conclusions are supported by observations made after the aponeurotic roof of the cubital tunnel was released in the current study and the nerve was maintained in situ. In this situation, intraneural pressure did not decrease significantly compared with the values noted before operative release at either the level of the cubital tunnel or four centimeters proximal to it.
NOTE: The authors thank A. Marc Tetro, M.D., and Thomas Pilgram, Ph.D., for technical assistance in performing this study.