Eight fresh nonpaired, previously frozen cadaver arms amputated at the midpoint of the humerus were used. The mean age and standard deviation of the donors at the time of death was 75 ± 6.6 years; there were four female and four male donors. Medical images were acquired with use of a sixty-four-slice computed tomography scanner (GE LightSpeed Ultra; General Electric, New Berlin, Wisconsin) (Fig. 1). To ensure accurate sectioning of the olecranon, a three-dimensional computer surface model was generated with use of custom-written software implemented within the Visualization Toolkit (VTK; Kitware, Clifton Park, New York). Cutting planes corresponding to the sequential levels of olecranon excision were calculated on the computed tomography surface model. The orientation of the cutting planes was determined with use of the greater sigmoid notch and the posterior aspect of the proximal part of the ulna. The plane containing the guiding ridge of the greater sigmoid notch was calculated with a least-squares algorithm, and the vector normal to this plane was oriented medial-lateral. Another plane was created with use of three points that were digitized on the posterior aspect, and the vector normal to this plane was oriented anterior-posterior. The cross product of the two vectors was oriented proximal-distal, and it served as the vector normal to the cut planes. The position of the first cut plane (0% resection) was fixed at the tip of the olecranon. The position of the last cut plane (100% resection) was fixed at the tip of the coronoid. The locations of the cut planes in between were defined by 12.5% increments between the first and last cut planes (Fig. 1). Physical surface landmarks on the posterior aspect of the proximal part of the ulna were digitized with use of a tracked stylus. The surface data from the digitized landmarks were then registered to the computed tomography surface model with use of the iterative closest-point least-squares algorithm5,6. This enabled the cutting planes in the computed tomography model to be directly mapped to the physical specimen, providing navigation to the resection process.
Specimens were thawed at room temperature (mean, 22° ± 2°C) for eighteen hours prior to testing. The tendons of the brachioradialis, biceps, brachialis, and triceps were sutured with use of a locking Krackow repair7. The sutures were attached to stainless steel cables, which were connected to pneumatic actuators and servomotors in an elbow simulator. A Steinmann pin was placed through the distal end of the radius and ulna to secure the forearm in neutral rotation. A second pin was placed through the third metacarpal, across the carpus, and into the distal end of the radius to fix the wrist in neutral flexion and extension.
Each specimen was mounted in an in vitro elbow motion simulator (Fig. 2)8. The simulator utilized computer-controlled pneumatic actuators and servomotors attached to tendons to simulate active elbow flexion in the varus, valgus, horizontal, and dependent (vertical) positions. The elbow motion simulator attempts to reproduce some of the complex forces and moments that occur in vivo with use of a cadaver model. The varus and valgus positions were chosen in the current study as relevant positions to evaluate the effect of olecranon excision on elbow stability. The varus gravity-loaded position placed the medial humeral epicondyle inferiorly and the lateral epicondyle superiorly. Similarly, the valgus gravity-loaded position placed the medial epicondyle superiorly and the lateral epicondyle inferiorly. Forces were applied to the tendons to simulate active flexion of the elbow from full extension. The muscle-loading protocol was based on the physiological cross-sectional area of the muscles and electromyographic data with use of an in vitro testing system previously developed and employed in our laboratory8,9. In addition to active testing, passive flexion was performed in the dependent, horizontal, varus, and valgus gravity-loaded positions. To produce passive flexion, the arms were gently advanced manually throughout their natural arc of motion until motion ceased. This was performed exclusively by one investigator. Motion was initiated by applying gentle pressure on the ulnar side of the fifth metacarpal. Kinematic data were obtained with use of the Flock of Birds electromagnetic tracking system (Ascension Technology, Burlington, Vermont) with a receiver affixed to the ulna and the transmitter referenced to the humerus. The system accuracy as reported by the manufacturer is 1.8 mm and 0.5° root-mean-squared deviation.
The posterior aspect of the proximal part of the ulna was exposed through a posterior midline longitudinal skin incision10. Superficial skin flaps were developed, and the subcutaneous border of the olecranon was exposed. Distally, the interval between the flexor carpi ulnaris and extensor carpi ulnaris was utilized. Care was taken to preserve all remaining capsuloligamentous structures. With use of a 0.4-mm oscillating saw, serial resections of the olecranon were performed at 12.5% increments with use of the aforementioned image-guided technique. To confirm that the cuts guided by the navigation were accurate, we digitized the surface of the cut olecranon and compared it with the three-dimensional model of the proximal part of the ulna generated from the computed tomography scan. A triceps advancement and repair to the remaining olecranon was then performed as described by Morrey11. The triceps sutures were passed through two transosseous tunnels and secured with use of a screw with two plastic washers on the posterior border of the ulna. The free suture ends were clamped between the washers to allow for repeated triceps repairs without the need to resuture the tendon for each serial resection. Kinematic data for intact specimens were obtained prior to sectioning of the olecranon.
Anatomical coordinate systems were generated for the ulna and the humerus from the locations of anatomical features, which were digitized with use of a tracked stylus on intact specimens with use of custom-written software implemented with the Visualization Toolkit. Kinematic data were calculated and transformed into these anatomical coordinate systems. Kinematic outcome variables included varus-valgus angulation and internal and external rotation of the ulna with respect to the humerus.
Statistical analysis was used as a three-way repeated-measures analysis of variance comparing flexion angle, olecranon resection, and active and passive motion (for both varus-valgus angulation and internal-external rotation). The intact specimens and each subsequent stage of resection were included in the analysis. Significance was defined at p < 0.05. A previous study that evaluated the impact of olecranon excision on elbow stability defined elbow stability as a percentage of the intact elbow for a given resection4. Thus, in addition to the analysis discussed above, we analyzed our data in a similar fashion to allow for comparison of our findings with this previous work.
Source of Funding
Funding for this study was provided by the Canadian Institute for Health Research.
There was no effect of flexion angle on mean varus-valgus angulation or rotation with the arm in the varus position under active motion (p > 0.05) (Fig. 3). With progressive resection of the olecranon, a progressive increase in varus angulation was observed (p < 0.04). Since there was no significant effect of flexion angle on varus-valgus angulation and rotation, only the mean data over the entire arc of flexion are reported throughout the results.
Varus-Valgus Angulation
Varus Position
Significant changes in elbow kinematics occurred with the arm in the varus orientation for both active (p < 0.04) and passive motion (p < 0.04) (Fig. 4). During active motion, progressive olecranon resection resulted in a significant increase in varus angulation. The mean varus angulation in the intact state was 2.8° ± 3.4°. This increased to 8.7° ± 6.8° as olecranon resection approached 100%. For passive motion, the mean angulation increased from 3.3° ± 3.3° in the intact state to 11.4° ± 11.3° with 100% resection. A significant increase in varus angulation was noted following the 0% olecranon excision and triceps repair compared with the intact elbow (p < 0.05). Resections of 87.5% and 100% were associated with visually detectable maltracking, and, in many cases, the elbows subluxated and the arc of motion was qualitatively observed to be reduced.
Valgus Position
Valgus angulation increased significantly with progressive resection of the olecranon with the arm in the valgus orientation (Fig. 4). Mean valgus angulation increased from 3.5° ± 4.1° in the intact state to 12.4° ± 3.6° when 100% of the olecranon was resected during active motion (p = 0.01). Mean valgus angulation for the intact specimens was 4.1° ± 5.2° and increased to 14.2° ± 3.1° after 100% of the olecranon was resected during passive motion (p = 0.001).
Dependent and Horizontal Positions
There was no significant difference in varus-valgus angulation following sequential excision of the olecranon with the arm in the dependent position for either active (p = 0.2) or passive (p = 0.2) motion. Similarly, no difference following olecranon excision was noted for active (p = 0.5) or passive (p = 0.2) motion with the simulator placed in the horizontal position (see Appendix).
Ulnohumeral Rotation
Varus Position
In the varus position, the degree of rotation significantly increased with progressive olecranon resections for both active (p < 0.001) and passive (p < 0.001) motion (Fig. 5). The ulna internally rotated relative to the distal end of the humerus as more olecranon was resected. With active motion, the mean external rotation of the intact elbow was 1.0° ± 7.8°. With progressive resections, the elbow underwent progressive internal rotation to a maximum of 13.4° ± 16.4° at 100% resection.
Valgus Position
With progressive olecranon resection, the amount of ulnar rotation relative to the distal end of the humerus increased significantly for both active (p < 0.007) and passive (p = 0.004) motion (Fig. 6). In the intact state, the ulna was oriented in a mean internal rotation of 0.5° ± 7.6° and externally rotated with progressive resection to a mean external rotation of 7.7° ± 11.0° with 100% resection. Passive motion demonstrated similar results. A mean of 0.2° ± 7.8° of internal rotation was measured for the intact state and increased to a mean of 7.9° ± 10.1° of external rotation with 100% resection.
Active Compared with Passive Kinematics
Varus-valgus angulation was significantly greater for passive compared with active motion with the simulator in the varus (p = 0.02) and valgus positions (p < 0.001). Rotation was also significantly greater for passive compared with active motion in the valgus position (p < 0.001). No difference in rotational kinematics was present in active or passive motion with the arm in the varus position (p = 0.1). Furthermore, no significant difference in rotational kinematics was noted in the dependent (p = 0.08) or horizontal (p = 0.8) simulator positions (see Appendix).
Elbow Stability: Comparison with the Intact State
For active valgus motion, relative elbow stability was maintained until 75% of the olecranon was excised (Fig. 7). Resections of up to 75% maintained greater than 70% of the stability of the intact elbow. Resection of greater than 75% resulted in an abrupt increase in instability.
Active varus motion also demonstrated small changes in elbow stability with resection of less than 37.5%. Resection of 37.5% to 50% caused the relative stability to decrease from greater than 70% to less than 50% of the intact state. A second pronounced decrease in stability was noted with resections of greater than 75%.
Elbow stability is dependent on both soft-tissue and osseous constraints. The coronoid, radial head, and ulnohumeral articulation have been identified as the primary osseous constraints of the elbow11,12-14. Recent biomechanical studies have evaluated the contribution of the radial head13,15-17 and coronoid11,18-21 to elbow stability. The olecranon is a key component of the ulnohumeral articulation. Despite this, there is a paucity of literature evaluating its contribution to elbow stability. We are aware of only two previous studies that have evaluated the contribution of the olecranon to elbow stability4,22. To our knowledge, the present study is the first to evaluate the role of the olecranon as an osseous elbow constraint with use of active and passive elbow kinematics.
In their landmark study, An et al. performed sequential resections (e.g., 25%, 50%, 75%, and 100%) of the olecranon in cadaver arms stripped of soft tissues4. The anterior capsule and collateral ligaments were left in place, and the arms were potted in aluminum cylinders. Varus and valgus loads were applied to the specimens with the arms in full extension or 90° of flexion, and varus-valgus angulation and rotation were measured with use of a materials testing machine. They reported a linear relationship between olecranon resection and displacement, and they concluded that 50% of the olecranon could be safely removed "[as] 60% of the constraining moment remained." On the basis of that study, it is generally considered that loss of up to 50% of the olecranon can be accepted in the management of olecranon fractures that are not amenable to open reduction and internal fixation. Although that study contributed to the understanding of the importance of the olecranon to elbow stability, testing was performed in only two static states, extension and 90° of flexion, with no information on dynamic elbow kinematics. Furthermore, the specimens were stripped of the surrounding musculature, which is now recognized as a key secondary elbow stabilizer.
A more recent study by Kamineni et al. evaluated the role of the posteromedial olecranon as a valgus restraint in throwing athletes22. They performed sequential resections of the posteromedial olecranon in 3-mm intervals in cadaver specimens. Specimens were passively moved through an arc of flexion. They concluded that valgus laxity increased as the posteromedial olecranon was excised. No significant difference in ulnohumeral rotation was noted.
The present study demonstrates that serial olecranon resection resulted in a progressive increase in varus-valgus angulation and ulnar rotation with the arm in both the varus and valgus orientations. This increase in joint laxity was observed throughout the range of motion and with both active and passive motion. Similar to the study by An et al.4, we found a relationship between the amount of olecranon resected and the degree of angular and rotational laxity. In fact, we noted significant changes in angular and rotational laxity with resection of as little as 12.5% of the olecranon (Figs. 4, 5, and 6). It therefore appears that elbow kinematics change significantly with even small resections of the olecranon, despite a secure triceps repair.
We did not observe that a 50% resection of the olecranon was a critical transition in elbow stability. Resections of 87.5% and 100% of the olecranon caused abrupt changes in elbow kinematics such that some elbows subluxated and were unable to be completely moved through a full arc of motion. Prior to 87.5% resection, the specimens demonstrated a nearly linear relationship between the magnitude of olecranon resection and the change in varus-valgus and rotational kinematics. It is likely that the marked changes in elbow kinematics observed with large olecranon excisions corresponded to the sectioning of the ulnar insertions of the anterior bundle of the medial collateral ligament and the lateral ulnar collateral ligament. The anterior bundle of the medial collateral ligament has been recognized as the primary restraint to valgus angulation12,18. With its sectioning and 87.5% of the olecranon resected, a sudden increase in elbow instability was observed. Our data suggest that preservation of the olecranon should be considered, if possible. However, up to 75% of the olecranon probably can be removed if reconstruction is not possible and the other primary and secondary constraints of the elbow are preserved.
As mentioned before, An et al. were the first to assess the impact of olecranon excision on overall elbow stability in vitro. They reported their results as relative elbow stability, expressing stability at each resection as a percentage of the intact state. We also analyzed our data in this format to allow for a comparison of findings between these two studies (Fig. 7). It must be noted that the data in our study are for active kinematics throughout a complete arc of motion, whereas the previous study was for passive motion in two static positions. Nonetheless, our results show a similar trend to those of An et al.4. As discussed above, abrupt changes in elbow kinematics occur at resections of greater than 75% for active valgus motion. At this point, (75% resection), the remaining olecranon with a secure triceps repair demonstrated 70% of the ability of the intact specimen to withstand varus-valgus angulation. This abruptly declined with resections of greater than 75%. Active motion in the varus position demonstrated two distinct transition points in stability compared with the intact specimen. The first occurred at 37.5% resection and the second, at 75% resection. Interestingly, smaller resections caused greater changes in relative elbow stability in the varus position compared with the valgus position. However, relative stability in both simulator positions demonstrated near catastrophic failure at 87.5% and greater amounts of resection. These findings suggest that alternative reconstruction techniques (i.e., bone-grafting) should be employed with unreconstructible olecranon defects of 87.5% or greater to avoid subluxation and dislocation.
We observed that even small amounts of olecranon resection resulted in an alteration of elbow kinematics. This may have resulted both from the loss of the osseous integrity of the ulnohumeral joint as well as from removal of the posterolateral joint capsule and the posterior bundle of the medial collateral ligament, which arise from the margins of the olecranon process. In a previous study from our laboratory, we found a mean increase in maximum varus-valgus laxity of 3.5° and maximum internal rotation of 1.0° following sectioning of the posterior bundle of the medial collateral ligament23. This suggests that while the increase in angulation following olecranon excision can likely be attributed to a loss of both the osseous and soft-tissue constraints, the osseous component is likely more important at least until large olecranon deficiencies, which disrupt the attachment of the anterior bundle of the medial collateral ligament, are present.
A difference in elbow kinematics was noted between passive and active range of motion. Varus and valgus angulation and ulnohumeral rotation were greater during passive range of motion compared with active range of motion. These differences are likely due to the simulated activation of the triceps, brachialis, biceps, and brachioradialis by the elbow testing system during active motion. On activation, the muscles apply a compressive force to the ulnohumeral joint, affording additional elbow stability. As our simulator used muscle loads based on muscle size and electromyographic activation data available in the literature, we believe our loading protocols likely physiological and therefore our observations likely are clinically relevant.
No difference in elbow kinematics was noted with olecranon resection in the dependent and horizontal simulator positions. In these positions, little change in varus-valgus angulation and ulnohumeral rotation occurred. These findings have implications for rehabilitation following operative intervention for olecranon fractures. In situations in which the olecranon is excised and a triceps advancement and repair is performed, our findings suggest that passive and active-assisted range of motion in dependent and horizontal positions can be initiated without substantially altering elbow kinematics. Passive motions with the arm in the varus and valgus orientations should be avoided, whereas active motion may be acceptable.
The present study is the first, to our knowledge, to evaluate the role of the olecranon while modeling active in vitro kinematics. Active motion allowed us to simulate in vivo elbow motion and more accurately model traumatic elbow pathology. Accurate quantification of three-dimensional unconstrained joint motion was achieved in the setting of a clinically relevant triceps advancement and repair. Furthermore, we utilized a novel tool to guide serial resections of the olecranon. Utilization of custom-written computer navigation surface-referencing software based on pretesting computed tomography scans allowed us to accurately plan resections without direct visualization of the articular surface. This also permitted the preservation of all of the soft-tissue attachments to the olecranon prior to each resection.
Our definition of the proximal part of the ulna differed from that in a previous study, in which the olecranon tip was chosen as the most proximal resection point4. In the present study, the most proximal aspect of the guiding ridge of the greater sigmoid notch was used. Our rationale for this designation was that it is the most proximal aspect of the articular surface. We assumed that the nonarticular portion of the proximal part of the ulna did not contribute to overall joint stability. Furthermore, the morphology of the tip of the olecranon is highly variable and provides a less reproducible landmark, despite its use in a previous study.
The present study has a number of limitations that should be emphasized. First, our model was designed to evaluate the contribution of the olecranon to elbow stability with intact soft tissues, coronoid, and radial head. Additional studies are needed to assess changes in elbow kinematics with soft-tissue deficiencies and in combination with osseous injuries. Second, this investigation was performed in cadaver specimens from elderly donors with use of an in vitro elbow-testing system, which likely does not precisely replicate the clinical situation. Finally, our study was not able to quantify the amount of varus-valgus and rotational angulation that would be clinically important. It is likely that low-demand elderly patients would tolerate larger amounts of olecranon excision than would high-demand young athletes. Clinical outcome studies are needed to confirm the observations of this in vitro biomechanical study.
The current study demonstrates that progressive loss of the olecranon results in a progressive increase in elbow instability as demonstrated by a significant increase in varus-valgus angulation and ulnohumeral rotation. Significant changes in kinematics were noted with intra-articular resections as small as 12.5%. While it is likely that symptomatic instability would not be evident with smaller magnitudes of olecranon loss, the long-term consequences of altered kinematics, such as degenerative arthritis and late instability due to failure of secondary restraints, require clinical correlation. The findings of this study support open reduction and internal fixation of displaced olecranon fractures, when technically possible, rather than excision and triceps advancement. Gross instability was observed with olecranon resections of 87.5% or greater. Thus, when open reduction and internal fixation of the olecranon is not possible in the setting of these large olecranon defects, this in vitro biomechanical study supports reconstruction with use of alternative techniques, including advancement of the olecranon tip, autografts, allografts, or prosthetic replacement, as they may yield superior results compared with olecranon resection and triceps advancement.
Finally, angulation and rotation with the elbow positioned in the horizontal and dependent positions did not differ significantly from intact elbows regardless of the magnitude of the olecranon resection. This study suggests that patients with larger olecranon defects should avoid arm positions demonstrated to evoke instability (i.e., varus and valgus) and limit rehabilitation to the horizontal and dependent positions.