Ten fresh-frozen cadaver arms, from donors with a mean age at the time of death of 66.3 years (range, forty-six to eighty-six years), six of whom were male, were amputated at the midhumeral level and were stored at -20°C. The arms were thawed at room temperature (22°C) for eighteen hours and then were prepared for mounting in an elbow-motion simulator. Sutures were secured to the distal end of the tendons of the wrist flexors (flexor carpi ulnaris and flexor carpi radialis), wrist extensors (extensor carpi ulnaris and carpi radialis longus), brachioradialis, pronator teres, supinator, biceps, brachialis, and triceps. The humerus was rigidly secured to the simulator with use of a clamp that allowed unconstrained elbow flexion15. Sutures were then passed through alignment guides and attached to stainless-steel cables. The alignment guide was placed at the medial epicondyle for the pronator teres and the wrist flexors, at the lateral epicondyle for the wrist extensors, and at the supracondylar ridge for the brachioradialis. The sutures, which attached to the triceps, biceps, and brachialis, were also directed through alignment guides to reproduce their physiologic line of action. To simulate the function of the supinator, a suture anchor was inserted into the radial tuberosity and the suture was passed through a plastic sleeve inserted in the radial aspect of the ulnar shaft. The suture then traveled through an intraosseous tunnel in the medullary canal and exited the proximal aspect of the olecranon. An in vitro elbow-motion simulator, utilizing computer-controlled pneumatic actuators and electronic motors attached to tendons, was used to simulate unconstrained active elbow flexion in the vertical (dependent) position. In this position, the long axis of the humerus is oriented perpendicular to the ground. To achieve varus and valgus gravitational loading, a hinge on the base plate of the testing apparatus allowed passive (manual) elbow flexion to be performed with the plane of flexion oriented horizontally (Fig. 1). In the varus position, the long axis of the humerus was parallel to the ground with the medial epicondyle facing down and the lateral epicondyle facing up. Conversely, in the valgus position, the long axis of the humerus was parallel to the ground, but the medial epicondyle was facing up and the lateral epicondyle was facing down. Varus-valgus and internal-external rotational kinematics were recorded with use of an electromagnetic tracking system16. The actuators applied specific forces to the tendons, first to position the forearm in supination or pronation, and then to move the elbow from full extension to full flexion at a constant controlled angular velocity (10 deg/s) (Table I). The muscle-loading protocol was based on electromyographic data and muscle cross-sectional area16,17.
Testing was first completed on the intact arm and was repeated after sectioning of the posterior medial collateral ligament and again after excision of the anterior capsule. Laterally, the extensor muscle mass was separated from the lateral collateral ligament and reflected off the lateral epicondyle. The lateral collateral ligament was then sectioned from its insertion on the lateral epicondyle. The lateral collateral ligament was repaired with use of a locking Krackow technique with number-2 Hi-Fi ultra-high molecular weight polyethylene suture (ConMed Linvatec, Largo, Florida). A 2-mm drill was used to make two diverging intraosseous tunnels. The first tunnel started on the central aspect of the capitellum at the center of axis of flexion of the elbow18 and exited the posterior aspect of the lateral supracondylar ridge. The second tunnel had the same starting point but exited the anterior aspect of the lateral supracondylar ridge. The ends of both sutures were passed through the bone tunnels, tied in a loop, and secured to a cable that was attached to a pneumatic actuator. The actuator was used to tension the lateral collateral ligament to 20 N, with the elbow flexed to 90° and the forearm in pronation. Once tensioned, the cable that attached to the lateral collateral ligament was secured to a clamp fixed to the base of the simulator. This lateral collateral ligament repair technique has previously been reported to effectively restore the kinematics and stability of the elbow similar to those of the intact joint19. The repair and tensioning techniques were consistent in all simulator positions (varus, valgus, and vertical). Testing was performed with the lateral collateral ligament released and tensioned in each position for each coronoid state.
Digital calipers (Digimatic CD-6; Mitutoyo, Tokyo, Japan) and an oscillating saw were used to sequentially simulate the anteromedial coronoid facet fractures (Fig. 2). For 2.5-mm subtype-I fractures, an osteotomy was completed from a point measured 2.5 mm posterior to the tip of the coronoid to a point measured 2.5 mm from the articular margin at the anterior edge of the sublime tubercle. An additional 2.5 mm was removed to simulate 5-mm subtype-I anteromedial coronoid fractures. For subtype-II anteromedial coronoid fractures, an osteotomy, parallel to the posterior flat spot on the olecranon, was performed 2.5 and 5 mm from the tip of the coronoid. Finally, to create a subtype-III fracture, an osteotomy, perpendicular to the posterior flat spot of the olecranon, was performed in the sagittal plane just lateral to the sublime tubercle.
The testing protocol was performed in the vertical (dependent) and varus and valgus gravity-loaded positions with the coronoid intact and with all anteromedial coronoid fracture subtypes. All six coronoid states were tested both with the lateral collateral ligament detached and after repair with use of the transosseous suture technique described above. Data were collected at each step in the testing protocol.
Receivers were secured to the ulna and radius with use of 3.5-mm bicortical screws. The position of the ulna and radius with respect to the humerus was recorded, in six degrees of freedom, with use of an electromagnetic tracking system (Flock of Birds; Ascension Technologies, Burlington, Vermont)15,16,20. Anatomical coordinate reference systems for each bone were established by digitizing appropriate osseous landmarks following the completion of testing21. Digitized landmarks included the capitellum, trochlear groove, midhumeral shaft, greater sigmoid notch, and ulnar styloid. Kinematic data were then transformed into these anatomical coordinate systems, permitting direct measurements of the motion of the ulna relative to the humerus (i.e., varus-valgus and rotational flexion pathways). Flexion angle, varus-valgus angle, and internal-external rotation of the ulna relative to the humerus were recorded in all testing positions. For the purposes of this study, instability was defined as an alteration in the varus-valgus angle and/or in internal-external rotation of the elbow relative to that of the intact native articulation.
Because of the anatomical differences between specimens, there were large variations in alignment and kinematics among the intact arms. The wide variance between specimens resulted in large standard deviations. For this reason, a repeated-measures analysis of variance and post hoc Student-Newman-Keuls tests were used for statistical analysis. This method compares all conditions for an individual specimen with the intact state of the same specimen. Statistical differences can be detected in the presence of large variances or standard deviations among subjects. The repeated-measures study design had sufficient power to detect a 3° difference among groups. Significance was defined as a p value of <0.05. The condition of the elbow and the position of forearm rotation were the two factors analyzed with the repeated-measures analysis of variance.
Source of Funding
Peer-reviewed funding for this project was provided by a research grant from the Canadian Institutes of Health Research.
Changes in varus-valgus angulation were most evident at 60° of elbow flexion, whereas internal-external rotational instability was more evident at lower flexion angles (30°). In light of this, for illustration purposes, the results obtained at 60° of flexion were selected to represent varus-valgus angulation, and the results at 30° of flexion were selected to represent internal-external rotation (Tables II and III). Although testing was completed after sectioning the posterior medial collateral ligament and after anterior capsule excision, there was no difference between these conditions and the intact elbows. Therefore, in order to simplify the results and focus on the effects of the anteromedial coronoid facet fracture subtypes, these data were not presented.
Vertical (Dependent) Position with Lateral Collateral Ligament Deficiency—Active Flexion
In the vertical position, when the lateral collateral ligament was deficient and the coronoid was intact, there was no significant difference in active kinematics compared with the intact elbow (p > 0.8) or compared with elbows with lateral collateral ligament deficiency combined with subtype-I and 2.5-mm subtype-II coronoid fractures (p > 0.9) (Figs. 3 and 4). Five-millimeter subtype-II fractures, with the forearm in pronation, demonstrated a mean (and standard deviation) of 4.8° ± 4.9° (p < 0.005) of varus angulation and a mean of 5.5° ± 5.3° (p = 0.1) of internal rotation compared with a mean of 6.5° ± 1.7° of valgus angulation and 1.6° ± 2.9° of external rotation in the intact elbow. Forearm supination restored angulation and rotation to those of the intact elbow (p > 0.9).
With the forearm in pronation, subtype-III fractures with lateral collateral ligament deficiency demonstrated increased varus angulation (mean, 10.2° ± 5.3°; p < 0.005) and internal rotation (mean, 9.2° ± 5.6°; p < 0.005) compared with the intact elbow (Figs. 3 and 4). Supination was able to restore angular (p = 0.4) and rotational (p = 0.9) stability to levels that were not significantly different from those of the intact elbow or the lateral collateral ligament-deficient elbow.
Vertical (Dependent) Position with the Lateral Collateral Ligament Repaired—Active Flexion
With active flexion in the vertical position, the kinematics of subtype-I and II anteromedial coronoid facet fractures with the lateral collateral ligament repaired were not significantly different from those of the intact elbow throughout flexion and irrespective of forearm position (p > 0.8) (Figs. 3 and 4). Although subtype-III fractures did not affect varus-valgus angulation (p > 0.9), internal rotation was increased (mean, 9.8° ± 5.5°; p < 0.005) compared with the intact elbow (mean, 1.6° ± 2.9° of external rotation) with the forearm in pronation. However, supination was able to restore varus-valgus and rotational stability to those of the intact elbow (p > 0.6) (Figs. 3 and 4).
Varus Position with Lateral Collateral Ligament Deficiency—Passive Flexion
With passive flexion in the varus position, there was severe instability when the lateral collateral ligament was deficient in the presence of an intact coronoid, with a mean of 18.6° ± 2.4° of varus angulation compared with a mean of 3.3° ± 1.5° of valgus angulation in the intact arm (p < 0.005). Varus instability increased with 2.5-mm (mean, 21.3° ± 3.1°; p < 0.005) and 5-mm (mean, 22.6° ± 2.6°; p < 0.005) subtype-I anteromedial coronoid fractures. This pattern of instability was even more evident with 2.5-mm subtype-II fractures (mean, 20.5° ± 2.7°; p < 0.005), with 5-mm subtype-II fractures (mean, 24.8° ± 2.6°; p < 0.005), and with subtype-III fractures (mean, 28.3° ± 2.9°; p < 0.005) (Figs. 5 and 6).
Varus Position with the Lateral Collateral Ligament Repaired—Passive Flexion
With passive flexion in the varus position, the kinematics of 2.5-mm subtype-I coronoid fractures with the lateral collateral ligament repaired were not significantly different from the intact elbow (p > 0.5) (Figs. 5 and 6). However, in pronation, 5-mm subtype-I fractures demonstrated a significant increase in internal rotation (mean, 6.2° ± 4.5° of internal rotation compared with a mean of 3.3° ± 3.1° of external rotation in the intact elbow; p < 0.05) but no change in varus angulation. In pronation, subtype-II 2.5-mm fractures with lateral collateral ligament repair demonstrated a mean of 0.3° ± 1.8° of varus angulation (p = 0.5) and 7.0° ± 4.5° of internal rotation (p < 0.005) (Figs. 5 and 6). The instability was increased with 5-mm subtype-II fractures (mean, 5.6° ± 2.0° of varus angulation [p < 0.005] and 10.8° ± 4.9° of internal rotation [p < 0.005]) and subtype-III fractures (p < 0.005) (Fig. 5).
Valgus Position with Lateral Collateral Ligament Deficiency—Passive Flexion
With passive flexion in the valgus position, when the lateral collateral ligament was deficient and the coronoid intact, there was no significant difference in active kinematics compared with the intact elbow (p > 0.8) or compared with elbows with lateral collateral ligament deficiency combined with subtype-I and 2.5-mm subtype-II anteromedial coronoid facet fractures (p > 0.7) (Fig. 7). In contrast, 5-mm subtype-II fractures demonstrated a mean of 14.9° ± 2.7° of valgus angulation (p < 0.005) and 3.6° ± 4.4° of internal rotation (p = 0.7) in pronation compared with a mean of 9.8° ± 1.6° of valgus angulation and 0.4° ± 3.2° of internal rotation in the intact elbow. Subtype-II and III fractures with lateral collateral ligament deficiency resulted in increased instability compared with isolated lateral collateral ligament deficiency (p < 0.05) (Fig. 7).
Valgus Position with the Lateral Collateral Ligament Repaired—Passive Flexion
With passive flexion in the valgus position, the kinematics of subtype-I and 2.5-mm subtype-II fractures with the lateral collateral ligament repaired were not significantly different from the intact elbow throughout flexion and irrespective of forearm position (p > 0.9) (Fig. 7). In contrast, 5-mm subtype-II fractures demonstrated increased valgus angulation (mean, 14.1° ± 2.5°; p < 0.05) but no significant change in internal rotation (mean, 4.1° ± 4.2°; p = 0.6) in pronation compared with a mean of 9.8° ± 1.6° of valgus angulation and a mean of 0.4° ± 3.4° of internal rotation in the intact elbow. Similarly, subtype-III coronoid fractures with the lateral collateral ligament repaired were significantly different from the intact elbow irrespective of forearm position (p < 0.005) (Fig. 7).
This study suggests that the size of the coronoid fracture fragment affects elbow kinematics, particularly with varus stress. The size of the anteromedial facet fracture and the presence of a concomitant lateral collateral ligament injury appear to be important determinants of the need for open reduction and internal fixation. This biomechanical study suggests that internal fixation of larger anteromedial coronoid facet fractures should be considered and that lateral collateral ligament repair alone should not be expected to restore kinematics in the majority of patients with this injury pattern; however, additional clinical studies are needed to determine patient outcomes and the effectiveness of internal fixation of these fractures. In contrast, when the lateral collateral ligament was deficient, small subtype-I anteromedial coronoid fractures (with an intact medial collateral ligament) demonstrated normal kinematics in the vertical (dependent) position. This implies that small subtype-I anteromedial coronoid fractures with an intact medial collateral ligament and a deficient lateral collateral ligament may be managed nonoperatively if strict rehabilitation protocols are followed to allow healing of the lateral collateral ligament. On the basis of this study and previous reports22, the recommended protocols include active and active-assisted range of motion with the forearm maintained in pronation and avoidance of shoulder abduction (varus stress). However, it is important to note that the medial collateral ligament was intact in this in vitro model. Clinically, these injuries can be associated with partial or complete injuries to the medial collateral ligament and the importance of evaluating both collateral ligaments cannot be overstated.
In contrast, although repair of the lateral collateral ligament was able to restore vertical (dependent) active kinematics for 5-mm subtype-I and 2.5-mm subtype-II anteromedial coronoid facet fractures, varus stability was not restored. There was significant varus and internal rotational instability in the varus gravity-loaded position. Since a secure anatomical repair of the lateral collateral ligament was performed and the medial collateral ligament was intact, this pattern of instability is likely related to the osteoarticular deficiency of the coronoid. Clinically, this is unlikely to improve over time. Larger (approximately 5-mm) subtype-I, subtype-II, and subtype-III fractures demonstrated significant varus and internal rotational instability even with the lateral collateral ligament repaired. These in vitro data suggest that internal fixation of these larger anteromedial coronoid facet fractures should be considered.
Subtype-II 5-mm anteromedial coronoid facet fractures with the lateral collateral ligament repaired demonstrated varus-valgus and rotational instability. Active flexion of the elbow in the vertical (dependent) position with the forearm in supination was able to restore kinematics similar to those of the intact arm. In contrast, passive supination resulted in increased varus and internal rotational instability with the arm oriented in both the varus and valgus gravity-loaded positions. When secure fixation of a subtype-II anteromedial coronoid facet fracture is not possible because of comminution, these in vitro results suggest that initial stability can be achieved with an intact medial collateral ligament, secure repair of the lateral collateral ligament, and a postoperative rehabilitation protocol consisting of active rather than passive motion of the elbow with the forearm maintained in supination. The fact that 5-mm subtype-II anteromedial coronoid facet fractures resulted in valgus instability supports previous studies that have demonstrated that larger coronoid fractures contribute to posterolateral rotatory instability23. The articulation of the coronoid tip with the lateral aspect of the trochlea likely provides both valgus and external rotatory stability.
Subtype-III coronoid fractures with and without lateral collateral ligament repair demonstrated severe varus and valgus instability, suggesting that open reduction and internal fixation of the coronoid and lateral collateral ligament repair are both necessary to restore stability. These fractures involve the sublime tubercle, which is the ulnar attachment of the medial collateral ligament, and result in valgus instability in addition to increased varus instability because of the osseous deficiency. The results of this study suggest that repair of the coronoid fracture should restore the function of the medial collateral ligament and correct the varus collapse into the osseous deficiency.
Varus instability caused by lateral collateral ligament deficiency was increased with a coexisting anteromedial coronoid fracture (all subtypes including 2.5-mm rim fractures). This indicates that, although isolated small (2.5-mm) rim fractures may not have a primary effect on varus instability, small fractures of the anteromedial facet rim can exacerbate varus instability in the setting of lateral collateral ligament deficiency (secondary effect).
This study did have limitations. The advanced age of the cadaver specimens could have affected the quality and integrity of the bone and soft tissues. Also, the stepwise experimental design made repeated loading of the specimens necessary. However, previous studies of dense connective tissue have suggested that the magnitude of deterioration in their biomechanical behavior during testing at room temperature is minor relative to the experimental effects of interest24. The size and location of the coronoid fractures simulated in our protocol may vary from the actual clinical fracture patterns. We did not study the effectiveness of anteromedial coronoid fixation because of the sequential resection protocol; additional studies are needed to evaluate the effectiveness of internal fixation. Finally, in vitro biomechanical studies have inherent limitations and the relationship of these changes to patient symptoms and the subsequent development of osteoarthritis require further study.
Despite these aforementioned weaknesses, our study had a number of strengths, including a test protocol that simulated motions as performed by patients in the postoperative period. Passive varus and valgus gravity-loaded positions and simulated active vertical (dependent) flexion were all examined. Also, since the mechanism of fracture-dislocations of the elbow includes a rotational component, rotational instability was quantified in all positions. Moreover, unlike previous biomechanical studies, transosseous repair of the lateral collateral ligament was used in this injury model to simulate the repair techniques performed clinically.
The results of this study suggest that isolated lateral collateral ligament repair can provide initial stability and normal kinematics with small displaced subtype-I anteromedial coronoid facet fractures with an intact medial collateral ligament. It may be possible to manage small subtype-I fractures without internal fixation if the medial collateral ligament is intact and the appropriate rehabilitation protocol is followed. However, clinical studies are needed to determine whether stability is maintained over time or whether the osseous defect will eventually lead to ligament attenuation and subsequent instability. Given that larger (=5-mm subtype-I, subtype-II, and subtype-III) fractures of the anteromedial coronoid facet resulted in varus and internal rotational instability, open reduction and internal fixation should help to restore stability. These recommendations are based on in vitro biomechanical data, and clinical studies are needed to determine whether nonoperative treatment of small subtype-I anteromedial coronoid facet fractures, with or without lateral collateral ligament repair, is sufficient to maintain stability and avoid the development of arthritis over time. 