Many disabling clinical conditions of the wrist, forearm, and elbow are treated by changing the relative length of the radius and the ulna or by excising portions of either bone. For example, the distal end of the ulna may be removed to reduce the incongruity of the joint after the deformity produced by a Colles fracture or the radius may be shortened (or the ulna may be lengthened) to alter load-bearing at the wrist in the treatment of avascular necrosis of the lunate (Kienböck disease)5,10. Most of these procedures are performed on the basis of a combination of historical and empirical success, but the rationale for these operations is based on limited relevant biomechanical data. Surprisingly little is known about the transmission of force from the hand to the elbow. The classic view has been that an axial compressive force applied to the palm is transmitted across the wrist to the distal end of the radius, through the interosseous membrane to the ulna, and finally across the elbow at the ulnotrochlear joint. Supporting evidence for this loading pathway is provided by the anatomical observation that the interosseous membrane has fibers that are aligned obliquely from the distal end of the ulna to the proximal end of the radius; theoretically, these fibers could transmit a longitudinal force from the radius to the ulna. Measurements of the force through the interosseous membrane have not been reported in the literature, to our knowledge.
Because of conflicting results from previous in vitro studies, it is not surprising that there is disagreement regarding the basic load-sharing mechanism in the forearm. The percentage of load borne by the ulna in the intact forearm has been reported to range from 9 to 43 per cent1-3,6-9,11. Carefully controlled studies of the transmission of load through the forearm are needed to help surgeons to understand the clinical success or failure of operative procedures on the forearm and to provide a biomechanical basis for operative decisions related to the relief of pain in the wrist.
The overall goal of the present study was to measure load-sharing of the radius and the ulna at the wrist and the elbow under a constant load applied to the wrist. We also sought to determine the role of the interosseous membrane in the transmission of load through the forearm. The objectives of the study were to determine (1) the effects of supination-pronation of the forearm, flexion of the elbow, and varus-valgus alignment of the elbow on this load-sharing and (2) the effects of shortening of the distal end of the radius on the pathways of load transmission at a single position of flexion and alignment of the elbow.
*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. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was Grant AR43735 from the National Institute of Arthritis and Musculoskeletal and Skin Diseases, the National Institutes of Health.
†Biomechanics Research Section, Department of Orthopaedic Surgery, Rehabilitation Center, University of California at Los Angeles, 1000 Veteran Avenue, Los Angeles, California 90095. E-mail address for Dr. Markolf: kmarkolf@ortho.medsch.ucla.edu.
‡535 North Cattail Way, Boise, Idaho 83704.
§Department of Orthopedic Surgery, Lenox Hill Hospital, 130 East 77th Street, New York, N.Y. 10021.
#The Hospital for Special Surgery, 535 East 70th Street, New York, N.Y. 10021.
Preparation of Specimens
Ten fresh-frozen forearms were obtained from the cadavera of individuals who had been fifty-six to seventy-eight years old at the time of death. Each specimen was sectioned in the middle of the humeral shaft, and all skin and muscle tissue was stripped away. The interosseous membrane, joint capsules at the wrist and elbow, and all ligaments were left intact. Anteroposterior and lateral radiographs were made to eliminate specimens with deformity. The first and fifth metacarpals were amputated at their bases, and the central three metacarpals were potted in a cylindrical mold with polymethylmethacrylate for mounting in the test fixture. The third metacarpal was aligned with the long axis of the cylinder during potting. The transected end of the humeral shaft was potted in a similar fashion, with its long axis aligned with the axis of the potting cylinder. To keep the specimen hydrated during testing, it was enclosed in a polyethylene bag, the interior surface of which had been moistened with Ringer solution.
Special miniature load-cells were designed and built for measurements of force in the radius and the ulna. The basic design consisted of a short, stiff beam element, 2.5 centimeters long, that was instrumented with strain-gauges. Two prongs, one extending from each end of the beam element, were cemented into the intramedullary canal of the distal end of the ulnar shaft and the proximal end of the radial shaft (Fig. 1-A). The load-cells were positioned (circumferentially) on the bone so that they did not impinge on the radius during a full range of supination and pronation of the forearm. With the prongs of the load-cells used as a guide, a transverse osteotomy was made in the proximal end of the radius and in the distal end of the ulna with use of a saw-blade. The compressive forces transmitted to the proximal end of the radius and the distal end of the ulna could then be measured by the beam elements of the load-cells, which also maintained anatomical alignment of the proximal and distal portions of each bone.
Loading Apparatus
Special loading fixtures were mounted on a servohydraulic testing machine (model 812; MTS, Minneapolis, Minnesota), which was operated under load control to apply a constant force of 134 newtons to the wrist (Fig. 1-B). The potted end of the humeral shaft was mounted on a test fixture that was attached to the crosshead of the testing machine. The angle of elbow flexion could be varied by tilting the fixture, and varus-valgus alignment of the elbow could be set by rotating the humerus within the clamping cylinder. Pivoting and sliding adjustments at the connection of the test fixture and the crosshead allowed for positioning of the forearm in a vertical orientation for all angles of elbow flexion. The potted metacarpals were clamped in a cup-like fixture, which was mounted on cross-slides, to position the wrist in correct anatomical alignment with respect to the distal ends of the radius and ulna. A thrust-bearing, mounted within a cylindrical housing beneath the cross-slides, allowed for manual rotation of the wrist (with minimum friction) under the constant load that was applied to the forearm. A rotary linear variable differential transformer mounted within the cylindrical housing measured rotation of the potted metacarpals, and the load-cell mounted on the hydraulic actuator of the testing machine recorded the load applied to the forearm. The cross-slide was adjusted and locked in order to position the wrist so that the distal end of the ulna was placed over the axis of the cylindrical housing containing the thrust-bearing. Because the anatomical axis of the forearm rotation passed through the distal end of the ulna, the distal end of the ulna remained stationary in space as the forearm was rotated. During testing, only rotation of the forearm was allowed as a constant axial load was applied through the wrist.
Test Procedure
The load-cells were calibrated before testing of each specimen. Bench calibrations of each load-cell with use of weights and a standardized loading fixture established linearity of the devices. However, those calibrations could not be used to determine force in the bones in situ because the moment arm from the line of action of the force in the bone to the beam element (that is, the effective prong length) was slightly different for each load-cell installation. Furthermore, the load-cells had a limited sensitivity to the bending moment that was produced by compression loading of a curved column. Therefore, separate load-cell calibrations were performed with both load-cells cemented into the bones. Because of the need for the separate calibrations, the load-cells were designed to allow detachment of the beam element from the cemented prongs after the unit had been installed.
Calibration of the proximal radial load-cell was a simple and straightforward procedure. The beam element of the distal ulnar load-cell was detached from the prongs, and the mobile distal segment of ulnar bone was tilted away from the rest of the ulna so that there was no osseous contact. This guaranteed that all axial load was transmitted through the radius (through the proximal radial load-cell). The radial load-cell then was calibrated by comparing its output with the load that was applied to the wrist.
Calibration of the distal ulnar load-cell was somewhat more involved. The beam element was removed from the prongs of the proximal radial load-cell and the proximal segment of radial bone was tilted away from the remainder of the radius so that no load could be transmitted to the capitellum. Two small metal retainers containing polymethylmethacrylate (in the doughy stage) were pressed against opposite sides of the ulnar head to make an impression of the osseous surface. When the cement had set, the retainers with the conforming molds were gripped in a yoke-like fixture attached to the crosshead of the materials testing machine. This allowed force to be applied directly to the clamped ulnar head and through the ulnar load-cell. The ulnar load-cell then was calibrated by comparing its output with the force applied to the (clamped) ulnar head, as recorded by the load-cell on the materials testing machine.
Once the load-cells had been calibrated, a standard test could be performed; the same basic procedure was used for each test. The ulna was always vertical during testing, with the wrist and the elbow adjusted to the desired positions. Neutral deviation and flexion of the wrist were attained by tilting the fixation cup containing the metacarpals. Tests were performed at three positions of elbow flexion: full extension, 45 degrees of flexion, and 90 degrees of flexion. The humerus was rotated and fixed within its cylindrical clamp so that the ulna was oriented vertically when a 1.0-newton-meter varus or valgus moment was applied to the forearm. As will be described, slight manual pressure was required to reduce the wrist and to maintain the forearm in a vertical position during testing. The axis of forearm rotation was aligned with the axis of the hydraulic actuator by adjusting the cross-slides until the distal end of the ulna was stationary in space as the wrist was rotated. Neutral rotation of the forearm was established by aligning the distal end of the radius and the distal end of the ulna in the flexion-extension plane of the elbow.
In the absence of contracted forearm muscles, a cadaveric wrist is a relatively loose and mobile articulation. Under compressive load, the proximal carpal row can readily subluxate volarly from its anatomical alignment with the distal ends of the radius and ulna. In addition, there is a certain degree of torsional laxity between the distal ends of the radius and ulna and the carpal row as the hand is rotated. In order to maintain anatomical alignment (alignment of the metacarpals, carpals, radius, and ulna in a single plane) during testing, the wrist was manually reduced with use of digital pressure as the distal end of the radius was rotated about the stationary ulna. In this manner, the carpal bones (and the potted metacarpals below them) were made to follow the radial head as a single unit as it was moved.
A typical test run began with the forearm in neutral rotation. A constant axial force of 134 newtons was applied to the potted metacarpals under load control, and the forearm was rotated manually, at approximately 5 degrees per second, to 60 degrees of supination and then to 60 degrees of pronation. The forces at the proximal and distal load-cells, the load applied to the wrist, and supination-pronation of the forearm were recorded continuously with use of a personal computer with an analog-digital board and data-acquisition software.
A specially designed slotted plate was fixed to the distal third of the radius for testing with radial shortening (Fig. 1-B). A segment of bone, approximately eight millimeters long, was removed from the section fixed with the plate, and the distance between the ends of the bone was measured with use of a dial caliper; this distance was the reference length for subsequent radial shortenings. The standard test was repeated with the plate in place to verify that no artifact had been introduced by application of the plate. The screws holding the plate to the bone then were loosened, the ends of the bone were brought together to shorten the radius, and the screws were retightened. The undersurface of the plate (the side in contact with the bone) was curved so as to conform closely to the osseous surface. The configuration of the plate and its slots helped to maintain alignment of the segments of the bone during shortening. The standard test was repeated, with the elbow in 45 degrees of flexion and valgus alignment, after two, four, and six millimeters of radial shortening.
Analysis of Data
For calculations of load-sharing, important simplifications were made regarding the pathway of load transmission between the radius and the ulna. We assumed that the loads transmitted between the radius and the ulna at their proximal and distal ligamentous interconnections were negligible compared with the force transmitted by the interosseous membrane, which is a larger structure with fibers oriented in a direction favorable for load transfer. Therefore, output from the proximal (radial) load-cell (which measured the contact force as well as the force transmitted from the radius to the ulna through the proximal radioulnar ligament) was assumed to represent only the contact force of the radial head. Similarly, output from the distal (ulnar) load-cell (which measured the contact force as well as the force transmitted from the radius to the ulna through the anterior and posterior ligaments of the distal radioulnar ligament complex) was assumed to represent only the contact force of the ulnar head. The fraction of the load that was transmitted to the forearm by articular contact at the proximal end of the ulna was calculated by subtracting the fraction that was registered by the radial load-cell from 1.00. The fraction of the load that was carried by the distal end of the radius was calculated by subtracting the fraction that was recorded by the ulnar load-cell from 1.00. The force through the interosseous membrane was calculated as the difference between the proximal and distal ulnar contact forces.
A two-way analysis-of-variance model with repeated measures was used to determine the significance of the mean differences between the forces recorded at the load-cells during testing under different status conditions and at discrete angles of forearm rotation. The angle of forearm rotation and the status conditions were fixed effects. The status conditions were varus and valgus alignment of the elbow, three positions of elbow flexion, and three fixed amounts of radial shortening. For analyses with only one independent variable (the angle of forearm rotation), a one-way analysis of variance with repeated measures was performed. Multiple pairwise comparisons were made with use of the Student-Newman-Keuls procedure. The level of significance was p = 0.05.
Previous Biomechanical Studies
A variety of methodologies have been used in experiments on cadavera to determine load-sharing in the forearm, and this has resulted in conflicting views on the transmission of load. Halls and Travill inserted one-millimeter-thick transducers that had been painted with a pressure-sensitive substance into the radiocapitellar and ulnotrochlear joints of formalin-preserved forearms from cadavera. They reported that 57 per cent of a 147-newton load applied to the hand was transmitted to the capitellum and 43 per cent was transmitted to the trochlea.
Palmer and Werner6 inserted miniature load-cells into the mid-portion of the radial and ulnar shafts of five forearms and applied load to the flexor and extensor tendons of the wrist with use of weights and pulleys. The forearm was sectioned just distal to the elbow, and the ends of the radius and the ulna were potted. All of the specimens were transfixed with a pin through the third metacarpal with the wrist in neutral deviation and neutral flexion, and testing was done only with the forearm in neutral rotation. They reported that 60 per cent of the load that was applied to the wrist was borne by the radius and 40 per cent was borne by the ulna6. In a later study, the authors repeated the experiments on sixteen specimens with the elbow joint intact7. The revised estimate of load-sharing was 82 per cent for the radius and 18 per cent for the ulna.
Using the same experimental approach to study the effects of flexion of the wrist and rotation of the forearm in nine specimens, af Ekenstam et al. reported that the load borne by the radius was somewhat less during flexion and full pronation than it was in the neutral position. The amount of load borne by the ulna ranged from 9 to 37 per cent.
Werner et al. used Palmer's basic loading methodology, with massive load-cells inserted into the distal ends of the radius and ulna, and reported that lengthening of the ulna by 2.5 millimeters increased the load borne by the ulna from 18.4 to 41.9 per cent. Shortening of the ulna by the same amount decreased the ulnar load to 4.3 per cent. These measurements were recorded with the forearm in neutral rotation only.
Using the same test methodology (but with the metacarpals potted), Trumble et al. repeated these experiments on ten specimens and reported that 17 per cent of the axial load was borne by the ulna. Ulnar load increased with extension of the wrist and ulnar deviation and decreased with flexion and radial deviation. Because there was no statistical analysis of the data, the results must be considered as observational only.
Hotchkiss et al. first reported that the central portion of the interosseous membrane was thickened and aligned in a manner that suggested that load could be transferred between the radius and the ulna. The central portion of the membrane functioned as a ligament and was found to be the primary stabilizer of the radius after excision of the radial head—that is, after radiocapitellar contact was lost3. The triangular fibrocartilage complex was shown to have a secondary role.
Rabinowitz et al. reported on the relative contributions of the interosseous membrane and the triangular fibrocartilage complex in preventing proximal migration of the radius after excision of the radial head. In their study of twelve specimens, they used a sagittal osteotomy of the humerus to divide the capitellum and trochlea and then mounted separate load transducers on each half. The specimens were tested with pronation of the forearm, extension of the wrist, and extension of the elbow (simulating a fall on an outstretched hand). Before excision of the radial head, 70 per cent of the load was transmitted to the capitellum and 30 per cent, to the trochlea. Those authors found that loss of both the central portion of the interosseous membrane and the triangular fibrocartilage complex was necessary for dramatic proximal translation of the radius.
Comparisons with Previous Studies
Several loading configurations have been used to apply force to the wrist of a cadaveric forearm. Palmer and Werner6,7 mounted the forearm in the vertical position (with the hand up) and suspended weights from the flexor and extensor tendons of the wrist. The weights were balanced to produce a zero net flexion-extension moment so that the metacarpals remained in a vertical orientation. With this loading scheme, load can be applied to the wrist with the forearm in neutral flexion-extension, neutral deviation, and neutral rotation, but rebalancing the weights to provide reproducible placements of the hand for supinated and pronated positions would be difficult with this method.
Rabinowitz et al. used a modification of this approach to apply load to the wrist in hyperextension. Weights were applied to transverse pins through the bases of the second and third metacarpals to forcibly extend the wrist. Only full extension of the wrist could be studied with this technique.
We chose the method of loading that was used by Trumble et al., who potted the metacarpals in a block of resin for subsequent application of load to the wrist. This technique allowed the hand to be positioned accurately for studies of pronation-supination of the forearm. The one modification that we have made to this loading scheme is to pot the middle three metacarpals only; the thumb and fifth metacarpal are considered to be of little importance in loading through the wrist.
There are three important ways in which our study differs from previous work. Most investigators have inserted load-cells either into the middle of the radial and ulnar shafts or into the distal one-third of those bones (at a substantial distance from the articular surface). With such placement, the sites of insertion of the interosseous membrane on both bones span the site of the load-cell. This configuration can provide data on load-sharing at one cross-section of the forearm only and cannot provide data on load that was transmitted to the radius and the ulna at locations of clinical interest (such as near the proximal and distal articular surfaces). Rabinowitz et al. mechanically divided the distal part of the humerus into trochlear and capitellar segments and installed a load-cell on each half. This configuration permitted measurement of load that was transmitted by articular contact at the trochlear and capitellar articulations at a single angle of elbow flexion.
In our study, force-transducers were inserted as close as possible to the ulnar and radial heads. These sites were chosen because of the relatively small diameters of the bones, which made central placement of the miniature load-cells easier. Because the total force that was applied to the wrist was also recorded, the proportions of load borne by the radius and the ulna could be calculated at both proximal and distal locations. Very little (if any) of the thin proximal or distal portion of the interosseous membrane spanned the site of either load-cell. The thickened central portion of the interosseous membrane is thought to be the main load-bearing portion of this structure.
The second area in which our study differed from earlier work relates to the effect of supination-pronation of the forearm on radioulnar load-sharing. Many grasping activities (such as twisting a doorknob) involve loading of the wrist with the forearm in a rotated position. Most previous investigators have performed all loading experiments with the specimen in a single rotational position (neutral or full pronation). In the one study of load-sharing in the forearm in which this variable was examined, af Ekenstam et al. measured only distal radioulnar load-sharing and the data were not analyzed statistically. Morrey et al. showed consistent increases in the force that was transmitted by the radial head when the forearm was in pronation, which is in agreement with our results.
The third, and perhaps most important, area that has been neglected in previous studies relates to the varus-valgus position of the elbow. At first glance, it may seem reasonable to assume that the radial head is always in direct contact with the capitellum during active contraction of the muscles of the forearm. However, there is a considerable degree of varus-valgus laxity in the flexed elbow, which allows lift-off of the radial head from the capitellum as a varus moment is applied. Although off-loading at this joint clearly could occur during certain dynamic loading of the forearm, it could be argued that non-contact at this point occurs commonly as the elbow is flexed and extended during daily activities. The study by Morrey et al. tends to support this possibility. They observed distraction of the radiocapitellar joint when the line of action of the brachial muscle pull was medial to the varus-valgus pivot point.
Interpretation of the Results
The articular contact force at the proximal end of the ulna was low at all positions of supination-pronation of the forearm when the elbow was in valgus alignment. This finding indicated that when there was initial contact of the radial head with the capitellum (as we established by applying a valgus moment to the elbow), virtually all of the load applied to the wrist was transmitted to the elbow directly through the radius. The output of the ulnar load-cell was slightly negative in some specimens, which indicated that the combination of ligament tension and osseous geometry acted to distract the ulnar head slightly rather than to compress it.
With varus alignment of the elbow, there was no initial articular contact at the radial head, and the mechanism by which load was transmitted through the ulna was markedly different from that with valgus alignment. The fraction of the load applied to the wrist that was registered by the ulnar load-cell remained relatively low, whereas the fraction of the load applied to the wrist that was borne by the proximal end of the ulna increased dramatically compared with that with valgus alignment. This meant that load was transferred from the radius to the ulna through the interosseous membrane. Such transfer of load occurred by proximal displacement of the radius relative to the ulna, which generated a tensile force in the interosseous membrane. The fraction of the applied load that was transferred through the interosseous membrane (which was equal to the difference between the contact forces for the proximal and distal ends of the ulna) was always lower with valgus alignment of the elbow than with varus alignment. There was generally less scatter in the data on the force in the interosseous membrane with valgus alignment than there was with varus alignment. To our knowledge, we are the first to report data on forces through the interosseous membrane.
One important biomechanical function of the interosseous membrane is to maintain a kinematic link between the radius and the ulna by tethering the bones together during pronation and supination. Our results suggest that, if radiocapitellar contact is always maintained, the interosseous membrane (and ulna) have little axial load-bearing function. However, if there is a gap at the radiocapitellar joint during flexion-extension activities, contraction of the muscles of the forearm acts to close the gap and to generate tension in the interosseous membrane. Our results suggest that, under these conditions, the interosseous membrane has an important axial load-bearing function in addition to a kinematic function. It is remarkable that this thin membrane is capable of performing both biomechanical functions.
Our findings regarding load-sharing with valgus alignment of the elbow do not agree with the often mentioned percentages, determined by Palmer and Werner7, of 80 per cent for radial load and 20 per cent for ulnar load. We found that approximately 3 per cent of the load applied to the wrist was borne by the distal end of the ulna when the elbow was in valgus alignment. This was confirmed in each specimen when the distal end of the ulna was grasped manually and its transverse mobility was tested while the forearm was under load. The ulnar head could be moved freely in each specimen, and it was readily apparent that the ulna was bearing little, if any, load.
In the normal wrist, there is a substantial gap between the distal end of the ulna and the proximal carpal row; the ulnar head has no direct contact with any bone that can transmit compression force. Our data suggest that when the wrist is in neutral anatomical alignment, the interposed triangular fibrocartilage complex is too compliant to transmit substantial compressive force directly to the ulnar head. With varus alignment of the elbow, the radius displaced proximally as the wrist was loaded, and it transferred load to the middle of the ulnar shaft through the interosseous membrane. Theoretically, this proximal displacement of the radius relative to the ulna could increase compression of the triangular fibrocartilage as well. This did occur, to some degree, with supination of the wrist (Fig. 4).
In our study, the load-cells were mounted proximal and distal to the sites of insertion of the interosseous membrane on the radius and the ulna. In the studies by Palmer and Werner6,7, the load-cells were mounted near the middle of the shafts of the bones, and the ulnar load-cell registered force that was transferred from the radius to the ulna through the interosseous membrane. We believe that Palmer and Werner found a higher percentage of load borne by the ulna because their study differed from ours with respect to the alignment of the elbow and the placement of the load-cells.
During the in situ calibrations of the load-cells, load was applied to each bone in straight compression. Under the test conditions in which the interosseous membrane was active (varus alignment), force that developed in the interosseous membrane could theoretically have produced a shear component and bending moment at the site of each load-cell. We estimate that, under a worst-case scenario (such as varus alignment with all shear force and bending moment produced by the interosseous membrane acting in the most sensitive planes of deformation), the error in force measured by the load-cell that was attributable to shear would be approximately 5 per cent and the error attributable to bending would be approximately 13 per cent. However, the forces that were transmitted through both load-cells were quite low when the elbow was in varus alignment (Fig. 4). With valgus alignment, the forces in the interosseous membrane were low and the in situ calibrations accurately represented actual loading conditions.
The results of our study are particularly interesting with respect to the amount of radial shortening that caused an alteration of load-sharing in the forearm. The site of the radial shortening, distal to the interosseous membrane, was important in our assessment of this effect. We performed tests with radial shortening only with the elbow in valgus alignment. Although the clinical practice of some surgeons is to shorten the radius by no less than three millimeters, our results show that significant changes in load-sharing can be produced by as little as two millimeters of radial shortening. The changes in load-sharing between two and four millimeters of radial shortening were greater (p = 0.05) than those between zero and two millimeters. We found that four millimeters of radial shortening produced approximately equal load-sharing between the radius and the ulna at the wrist with the elbow in valgus alignment. However, the force in the interosseous membrane was unchanged under this condition. In fact, the data suggest a trend toward a slight transfer of load from the ulna to the radius with radial shortening (Fig. 8). The clinical consequences of this are unknown.
This study was performed on specimens from an elderly population. Because the anatomy of these forearms was radiographically normal, the effects of varus-valgus alignment and radial shortening (the two most important findings of the study) should apply to a younger population as well. Although great care was taken in calibrating the load-cells, some elastic deflection of the load-cell (closing of the gap in the bone) occurred under load. In the present study, load was generated in the forearm by direct compression through the wrist. During gripping activities, compressive load across the wrist is generated by muscles in the forearm that also act to stabilize the wrist. During testing, anatomical alignment of the wrist was maintained by manual pressure during rotation of the forearm so that the carpals, hand, and distal end of the radius moved together as a unit. This did not alter the compressive force measured by the load-cell because the manual forces that were applied to reduce the wrist were acting transverse to the direction in which the load was applied.
Our tests were performed with application of only 134 newtons of load to the wrist; this was the maximum load that we could apply without causing structural failure of the distal end of the ulna at the site of the slot for insertion of the load-cell. The cortical bone at this location was often very thin and weak. The forearm may bear higher loads in vivo, especially during strong gripping activities or when the forearm is used for upper-body support. We could not expect greater loads applied to the wrist to substantially change the radioulnar distribution of load with valgus alignment because radiocapitellar contact had already been established. With varus alignment, the proportions of the load borne by the radius and the ulna depend on the force in the interosseous membrane, which in turn depends on the amount of proximal displacement of the radius relative to the ulna before radiocapitellar contact has been established. At the present time, the amount of in vivo force that is applied to the forearm through the wrist during these activities is unknown.
NOTE: The authors gratefully acknowledge the assistance of Chi Rey Chen, who helped with the testing during the initial phases of the study.