the radial collateral ligament, the lateral ulnar collateral ligament, and the annular ligament 13

1823 COPYRIGHT 2001 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED Ligamentous Stabilizers Against Posterolateral Rotatory Instability of the Elbow BY CYNTHIA E. DUNNING, PHD, ZANE D.S. ZARZOUR, MD, STUART D. PATTERSON, MBCHB, JAMES A. JOHNSON, PHD, AND GRAHAM J.W. KING, MD Investigation performed at the Bioengineering Research Laboratory, Lawson Health Research Institute, Hand and Upper Limb Centre, St. Joseph s Health Care London, London, Ontario, Canada Background: The lateral ulnar collateral ligament, the entire lateral collateral ligament complex, and the overlying extensor muscles have all been suggested as key stabilizers against posterolateral rotatory instability of the elbow. The purpose of this investigation was to determine whether either an intact radial collateral ligament alone or an intact lateral ulnar collateral ligament alone is sufficient to prevent posterolateral rotatory instability when the annular ligament is intact. Methods: Sequential sectioning of the radial collateral and lateral ulnar collateral ligaments was performed in twelve fresh-frozen cadaveric upper extremities. At each stage of the sectioning protocol, a pivot shift test was performed with the arm in a vertical position. Passive elbow flexion was performed with the forearm maintained in either pronation or supination and the arm in the varus and valgus gravity-loaded orientations. An electromagnetic tracking device was used to quantify the internal-external rotation and varus-valgus angulation of the ulna with respect to the humerus. Results: Compared with the intact elbow, no differences in the magnitude of internal-external rotation or maximum varus-valgus laxity of the ulna were detected with only the radial collateral or lateral ulnar collateral ligament intact (p > 0.05). However, once the entire lateral collateral ligament was transected, significant increases in internalexternal rotation (p = 0.0007) and maximum varus-valgus laxity (p < 0.0001) were measured. None of the pivot shift tests had a clinically positive result until the entire lateral collateral ligament was sectioned. Conclusions: This study suggests that, when the annular ligament is intact, either the radial collateral ligament or the lateral ulnar collateral ligament can be transected without inducing posterolateral rotatory instability of the elbow. Clinical Relevance: Surgical approaches to the lateral side of the elbow that violate only the anterior or posterior half of the lateral collateral ligament should not result in posterolateral rotatory instability of the elbow. This is important information for surgeons planning various procedures on the lateral aspect of the elbow, such as reconstruction of a fractured radial head, radial head replacement, or total elbow arthroplasty. Posterolateral rotatory instability of the elbow is a term used to describe rotatory subluxation of the ulna relative to the humeral trochlea together with posterolateral dislocation of the radial head relative to the capitellum, without associated instability of the proximal radioulnar joint 1,2. This condition can be diagnosed clinically through the use of a lateral pivot shift test 1-3. Experiments have shown posterolateral rotatory instability to involve approximately 15 to 30 of external rotatory subluxation of the ulnohumeral joint 3,4. Injury to the lateral ulnar collateral ligament has been described as the chief cause of posterolateral rotatory instability 1,2,4-7. More recently, however, it has been suggested that the entire lateral collateral ligament complex is important for varus joint stability and that, although the lateral ulnar collateral ligament contributes to stability, it is not the sole constraint against posterolateral rotatory instability of the elbow 8-12. The lateral collateral ligament complex is composed of the radial collateral ligament, the lateral ulnar collateral ligament, and the annular ligament 13 (Fig. 1). The annular ligament inserts on the anterior and posterior margins of the lesser sigmoid notch of the ulna, passing over the radial head. Its primary function is to stabilize the proximal radioulnar joint 13,14. Since the condition of posterolateral rotatory instability specifically indicates that this joint is stable, we believe that the annular ligament is usually intact in a patient with this disorder. Therefore, this investigation deals only with the radial collateral and lateral ulnar collateral ligaments, which we refer to collectively as the lateral collateral ligament (Fig. 1). The purpose of this in vitro investigation was to determine whether either an intact radial collateral ligament alone or an intact lateral ulnar collateral ligament alone is sufficient to prevent posterolateral rotatory instability of the elbow when the annular ligament is intact. We hypothesized that both the radial collateral and the lateral ulnar collateral liga-

1824 Fig. 1 The lateral collateral ligament complex consists of the radial collateral ligament (RCL), the lateral ulnar collateral ligament (LUCL), and the annular ligament (AL). The lateral collateral ligament (LCL) consists of the radial collateral ligament and the lateral ulnar collateral ligament. The dashed line indicates the dividing line used to distinguish the radial collateral ligament from the lateral ulnar collateral ligament. ments are important stabilizers against posterolateral rotatory instability of the elbow. Materials and Methods welve fresh-frozen cadaveric upper extremities (mean T age [and standard deviation] of the individuals at the time of death, 72 ± 14 years; range, forty-seven to ninety-three years), stored at 20 C, were thawed overnight prior to testing. None of the specimens showed evidence of ligamentous injury or osteoarthritis. The humerus was transected at the midhumeral level, and all soft tissue 10 cm proximal to the elbow joint was removed. This allowed the specimen to be secured into a humeral mounting clamp affixed to a polyethylene baseplate. The base-plate was on a two-degrees-of-freedom hinge, permitting orientation of the arm in the varus or valgus gravityloaded positions as well as in the vertical position (Fig. 2). With the arm in the valgus gravity-loaded orientation, a single investigator moved the elbow passively through a full range of flexion while maintaining the forearm in either pronation or supination. These motions were repeated with the arm in the varus orientation, in which gravity provided a provocative varus stress to the elbow without the need to apply additional external loads. With the arm in the vertical position, a pivot shift test was performed by a single clinician familiar with the maneuver (G.J.W.K.). Testing was first performed with all skin and soft-tissue structures intact. Next, a surgical approach was made on the lateral side of the elbow through the Kocher interval 15. The common extensor muscles were identified and were elevated off the lateral epicondyle and the underlying lateral collateral ligament complex. The anconeus muscle was reflected posteriorly to expose the underlying ligaments and joint capsule. The anterior and posterior aspects of the capsule were sectioned as far medial as the coronoid and olecranon processes, respectively. Osseous landmarks were used to ensure that the sectioning of the capsule was reproducible among specimens. In all cases, the anteriormost border of the radial collateral ligament was defined as the tissue that passed from the lateral epicondyle to the annular ligament at the anterior margin of the radial head. With the lateral collateral ligament complex exposed, the supinator crest was identified. A marker was used to draw a line from the proximal edge of the crest to the posterior margin of the lateral collateral ligament at its insertion on the lateral epicondyle. A second line, drawn from the distal edge of the crest to the midpoint of the lateral epicondyle, served as the dividing line between the radial collateral ligament anteriorly and the lateral ulnar collateral ligament posteriorly (Fig. 1). In six of the specimens, the lateral ulnar collateral ligament was then sectioned at its insertion on the lateral epicondyle to model the location of ligament disruption observed clinically 16 ; the radial collateral ligament was left intact. The fascia between the extensor carpi ulnaris and the anconeus was repaired with use of number-1 braided absorbable suture, and the passive motion testing protocol was repeated. Following this, the Kocher interval was reopened and the radial collateral ligament was sec- Fig. 2 A humeral clamp was used to rigidly fix each specimen to a polyethylene base-plate, which could be placed in the varus (A), valgus (B), and vertical (C) positions. Motion of the ulna relative to the humerus was recorded by an electromagnetic tracking device, with the transmitter secured to the base-plate and a receiver fixed to the ulna.

1825 TABLE I Maximum Varus-Valgus Laxity and Corresponding Flexion Angle During a Pivot Shift Test Ligament State Forearm Position Maximum Varus-Valgus Laxity* (deg) Flexion Angle of Maximum Varus-Valgus Laxity* (deg) Intact (n = 12) Pronated 8.3 ± 3.3 77.8 ± 42.0 Supinated 7.7 ± 3.1 80.4 ± 44.4 Radial collateral ligament intact (n = 6) Pronated 13.6 ± 4.1 115.8 ± 29.2 Supinated 13.3 ± 4.7 117.5 ± 29.8 Lateral ulnar collateral ligament intact (n = 6) Pronated 8.4 ± 2.4 50.8 ± 47.7 Supinated 7.5 ± 2.1 68.3 ± 46.2 Lateral collateral ligament cut (n = 12) Pronated 37.5 ± 12.3 103.8 ± 13.3 Supinated 41.8 ± 12.4 102.1 ± 12.5 *The values are given as the mean and the standard deviation. tioned from the lateral epicondyle, so that the entire lateral collateral ligament between the ulna and the humerus was divided while the annular ligament remained intact. The Kocher interval was again sutured closed, and the passive elbow motions were repeated. In the other six specimens, a similar surgical protocol was followed, but the radial collateral ligament was sectioned prior to the lateral ulnar collateral ligament. An electromagnetic tracking device (Flock of Birds; Ascension Technology, Burlington, Vermont) was used to determine the orientation of the ulna relative to the humerus throughout the testing protocol 17. The device s transmitter was rigidly fixed to the polyethylene base-plate, effectively eliminating relative motion between the humerus and the transmitter. A single receiver was secured to the medial side of the distal part of the ulna so that it did not limit forearm rotation. Care was taken to ensure that the distance between the receiver and the transmitter always remained within the optimum operating range of the tracking system 17. During testing, the position and orientation of the receiver relative to the transmitter was recorded in six degrees of freedom. Osseous landmark digitization of several points on the humerus and ulna, as previously described 18, was performed at the completion of testing with use of a stylus probe attached to a second receiver. This enabled the spatial relationships between the humerus and the transmitter, and between the ulna and its receiver, to be determined. Therefore, the receiver with respect to transmitter data collected during testing could be converted to ulna with respect to humerus data during post hoc data analysis. Motion data obtained with the arm in the varus and valgus gravity-loaded positions were used to calculate the maximum varus-valgus laxity of the arm throughout the arc of flexion 18. This involved determining the varus angulation of the elbow with the arm in the varus orientation and the varus angulation of the elbow with the arm in the valgus orientation, at every 5 of flexion. Varus-valgus laxity was defined as the difference between these values (i.e., varus orientation angle minus valgus orientation angle). Thus, varus-valgus laxity was calculated at every 5 of flexion, and the largest value was recorded as the maximum varus-valgus laxity. The flexion angle at which the maximum varus-valgus laxity occurred was also noted. Data generated from the pivot shift test were used to determine the internal-external rotation of the ulna relative to the humerus, as calculated with use of an Euler Z-Y-X analysis. These results were used as a measure of rotational instability of the elbow. Statistical comparisons were performed with use of two-way repeated-measures analyses of variance and Student- Newman-Keuls multiple comparison procedures with alpha set at 0.05. Results nalysis of the maximum varus-valgus laxity data indicated A that, regardless of whether the radial collateral ligament or lateral ulnar collateral ligament was sectioned first, there was no difference in stability between the intact specimens and the specimens with either the radial collateral ligament (p > 0.05) or the lateral ulnar collateral ligament intact (p > 0.05) (Fig. 3). However, once both components of the lateral collateral ligament were sectioned, there was a marked increase in the maximum varus-valgus laxity (p < 0.0001). Regardless of the cutting sequence, forearm pronation or supination had no effect (p > 0.23). The angles at which the maximum varus-valgus laxities occurred were highly variable (Table I). The magnitudes of internal-external rotation during the pivot shift test are shown in Figure 4. The rotational stability of the intact specimens was not significantly different from that of the specimens with either the radial collateral ligament (p > 0.05) or the lateral ulnar collateral ligament intact (p > 0.05). When the entire lateral collateral ligament was cut, however, the magnitude of internal-external rotation increased significantly (p = 0.0007). None of the twelve specimens had a positive pivot shift test, as determined qualitatively by the clinician who performed the maneuver, until the entire lateral collateral ligament was

1826 sectioned. This finding was supported by the quantitative data, as the average magnitude of ulnar external rotation did not exceed 6 prior to transection of the entire lateral collateral ligament (Fig. 4). To determine the repeatability of the data generated from the manual pivot shift test, this maneuver was repeated ten times at each stage of the sectioning protocol in one of the specimens tested. The standard deviation in the internalexternal rotation data for the ten trials was calculated at every 5 of flexion. The maximum standard deviations were found to be 0.3 (at 50 of flexion) with the lateral collateral ligament intact, 0.4 (at 70 of flexion) after sectioning of the radial collateral ligament, and 2.6 (at 110 of flexion) after transection of the entire lateral collateral ligament. Fig. 3 The means (and standard deviations) of the maximum varus-valgus laxity measured during passive elbow flexion with the forearm maintained in either pronation or supination and the arm placed in the varus or valgus gravity-loaded orientation are shown. Regardless of the cutting sequence, no difference in maximum varus-valgus laxity was measured, compared with that in the intact state, until the entire lateral collateral ligament was cut. Fig. 4 The means (and standard deviations) of internal-external rotation measured during the pivot shift test with the arm in a vertical orientation are shown. Positive values indicate external rotation of the ulna. Regardless of the cutting sequence, no difference in internal-external rotation was measured, compared with that in the intact state, until the entire lateral collateral ligament was cut. Discussion hen O Driscoll et al. first de- posterolateral rotatory in- Wscribed stability of the elbow, they surmised that laxity of the lateral ulnar collateral ligament was the cause of the condition 1,2, and this hypothesis was supported by their subsequent anatomic study of the lateral ulnar collateral ligament 7. Our study demonstrated that, provided the radial collateral and annular ligaments remain intact, sectioning of the lateral ulnar collateral ligament and joint capsule does not result in posterolateral rotatory instability. This observation is in agreement with the recent finding by Hannouche and Begue that isolated sectioning of what they referred to as the medial bundle of the lateral collateral ligament (i.e., the lateral ulnar collateral ligament) resulted in only minor elbow laxity 19. However, Hannouche and Begue reported that sectioning of both the medial and the anterior bundle (i.e., the lateral ulnar collateral and radial collateral ligaments) resulted in ulnohumeral subluxation. Cohen and Hastings 8, Imatani et al. 9, and Olsen et al. 10-12 also speculated that the lateral ulnar collateral ligament is not the only constraint against posterolateral rotatory instability of the elbow; they believed that the entire lateral collateral ligament complex and surrounding tissues are important stabilizers.

1827 When the lateral collateral ligament was exposed, we could not visually establish, in a reproducible manner, a distinction between the radial collateral and lateral ulnar collateral ligaments near their humeral insertions. Olsen et al. commented that distinct separation of the lateral ulnar collateral ligament fibers within the lateral collateral ligament could not be accomplished proximal to the annular ligament 10. In our study, we defined the radial collateral and lateral ulnar collateral ligaments on the basis of osseous landmarks, by dividing the attachment site on the lateral epicondyle into anterior and posterior halves. This definition may also be useful clinically. We believe that by using the supinator crest of the ulna and the lateral epicondyle of the humerus to define the lateral ulnar collateral ligament we were able to reproducibly section this ligament from its origin on the epicondyle, the typical site of clinical disruption 16. Olsen et al. performed a series of studies to investigate the kinematics of the lateral ligamentous structures of the elbow 10-12. Although they sectioned the lateral ulnar collateral ligament distally, at its ulnar insertion, they too found that transection of the joint capsule and the lateral ulnar collateral ligament did not produce marked laxity 11. The results of our study are in agreement with their finding that varus-valgus laxity and internal-external rotation of the ulna increased dramatically after the entire lateral collateral ligament was detached from its origin on the lateral epicondyle 10-12. Cohen and Hastings reported that the muscles and fascia overlying the lateral collateral ligament provide an important stabilizing influence against rotatory displacement in the elbow 8. The sectioning protocol employed in our study did not allow the stabilizing influence of this overlying tissue to be quantified. The surgical approach through the Kocher interval sacrificed the humeral origin of the common extensor muscles. Although the muscle origin was not reattached, the fascial incision made through the Kocher interval was sutured closed after each phase of the sectioning protocol. The elbow appeared to be more stable once this incision was closed, and we agree with Cohen and Hastings that this interval should be repaired following lateral elbow surgery. In our study, no specimen had a positive pivot shift test until the entire lateral collateral ligament was sectioned. In a study designed to investigate the lateral elbow stabilizers, Sojbjerg et al. concluded that the lateral collateral ligament had a minor stabilizing role and the annular ligament was the primary stabilizer 20. We did not assess the role of the annular ligament, which we believe is usually intact in patients with posterolateral rotatory instability of the elbow 1,2. Furthermore, in the protocol of Sojbjerg et al., the lateral collateral ligament was always sectioned following radial head excision and transection of the annular ligament. This may help to explain the differences between their results and ours. Comparison of the data for the intact radial collateral ligament with those for the intact lateral ulnar collateral ligament, as shown in Figures 3 and 4, reveals an interesting trend. These data suggest that the lateral ulnar collateral ligament may afford somewhat greater varus-valgus and rotational stability than the radial collateral ligament does. This finding appears to support the results of previous studies indicating the importance of the lateral ulnar collateral ligament in preventing posterolateral rotatory instability of the elbow 1,2,4,7. However, our study was not designed to directly compare the stability afforded by the radial collateral ligament with that afforded by the lateral ulnar collateral ligament, so the sample size was not large enough to test this trend statistically. A post hoc analysis of the pivot shift data revealed that the power to detect a 5 difference between the rotational stability afforded by the lateral ulnar collateral ligament and that afforded by the radial collateral ligament was only 20% and that a sample size of sixty would be required to detect this difference with a power of 80%. O Driscoll et al. found that posterior rotatory instability of the elbow involves approximately 15 of external rotatory subluxation 4, and our findings concur. None of the pivot shift tests in our study were deemed to be clinically positive nor did the magnitudes of external ulnar rotation exceed 15 until the entire lateral collateral ligament was sectioned. O Driscoll et al. also noted that, during the pivot shift test, the ulna undergoes maximum posterolateral rotatory displacement at approximately 40 of flexion and additional flexion reduces the joint 1,2. This finding does not agree with our data, which indicated that the ulna continues to externally rotate until 90 of elbow flexion (Fig. 4). This observation corresponds closely to the data of Olsen et al., who also demonstrated that external rotation of the ulna relative to the humerus increased until 90 to 100 of elbow flexion during a pivot shift test after sectioning of the lateral collateral ligament 12. It is interesting to note, however, that the pivot shift tests that we performed after transection of the entire lateral collateral ligament would still have been deemed to be clinically positive by 40 of flexion, since the mean external rotation of the ulna at this point was 17. A potential limitation of the present study stems from the use of the pivot shift test, which is a manually applied test that can be difficult to perform. It involves moving the elbow through an arc of flexion while simultaneously applying loads of axial compression, valgus moment, and forced supination 1,6. Therefore, the ability to reproduce this maneuver accurately after each stage of ligament sectioning in twelve specimens might be questioned. However, all of the pivot shift tests were performed by the senior author, who has considerable experience with the maneuver. In the specimen that was subjected to the pivot shift test ten times, the results were repeatable. Furthermore, the quantitative results reflect the instability that was observed clinically at each stage of the sectioning protocol. Our results suggest that both the radial collateral ligament and the lateral ulnar collateral ligament have a key role in the prevention of posterolateral rotatory instability of the elbow. This is important information for surgeons planning a lateral surgical approach to the elbow for a procedure such as a fracture repair or an arthroplasty 15. It suggests that either, but not both, of these structures can be sectioned to gain access to the elbow joint if the annular ligament is intact. If ei-

1828 ther the radial collateral or the lateral ulnar collateral ligament remains intact, there should be little risk of posterolateral rotatory instability developing postoperatively. If one portion of the lateral collateral ligament is transected, it is still recommended that it be repaired during surgical closure. NOTE: The authors thank Ms. Teresa Duck for her assistance in data collection. Cynthia E. Dunning, PhD Zane D.S. Zarzour, MD James A. Johnson, PhD Departments of Mechanical and Materials Engineering (C.E.D., J.A.J., and G.J.W.K.), Surgery (Z.D.S.Z., J.A.J., and G.J.W.K.), and Medical Biophysics (J.A.J. and G.J.W.K.), The University of Western Ontario, 1151 Richmond Street, Suite 2, London, ON N6A 4B8, Canada Graham J.W. King, MD Bioengineering Research Laboratory, Lawson Health Research Institute, Hand and Upper Limb Centre, St. Joseph s Health Care London, 268 Grosvenor Street, London, ON N6A 4L6, Canada. E-mail address: gking@uwo.ca Stuart D. Patterson, MBChB The Bond Clinic, 500 East Central Avenue, Winter Haven, FL 33880-3094 In support of their research or preparation of this manuscript, one or more of the authors received grants or outside funding from the Medical Research Council of Canada. None of the authors received payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated. References 1. O Driscoll SW, Bell DF, Morrey BF. Posterolateral rotatory instability of the elbow. J Bone Joint Surg Am. 1991;73:440-6. 2. O Driscoll SW, Morrey BF, Korinek SL, An KN. The pathoanatomy and kinematics of posterolateral rotatory instability (pivot-shift) of the elbow. Trans Orthop Res Soc. 1990;15:6. 3. Lee ML, Rosenwasser MP. Chronic elbow instability. Orthop Clin North Am. 1999;30:81-9. 4. O Driscoll SW, Morrey BF, Korinek S, An KN. Elbow subluxation and dislocation. A spectrum of instability. Clin Orthop. 1992;280:186-97. 5. Morrey BF, O Driscoll SW. Lateral collateral ligament injury. In: Morrey BF, editor. The elbow and its disorders. 2nd ed. Philadelphia: WB Saunders; 1993. p 573-80. 6. Nestor BJ, O Driscoll SW, Morrey BF. Ligamentous reconstruction for posterolateral rotatory instability of the elbow. J Bone Joint Surg Am. 1992;74:1235-41. 7. O Driscoll SW, Horii E, Morrey BF, Carmichael SW. Anatomy of the ulnar part of the lateral collateral ligament of the elbow. Clin Anat. 1992;5:296-303. 8. Cohen MS, Hastings H 2nd. Rotatory instability of the elbow. The anatomy and role of the lateral stabilizers. J Bone Joint Surg Am. 1997;79:225-33. 9. Imatani J, Ogura T, Morito Y, Hashizume H, Inoue H. Anatomic and histologic studies of the lateral collateral ligament complex of the elbow joint. J Shoulder Elbow Surg. 1999;8:625-7. 10. Olsen BS, Vaesel MT, Helmig P, Sojbjerg JO, Sneppen O. Lateral collateral ligament of the elbow joint: anatomy and kinematics. J Shoulder Elbow Surg. 1996;5:103-12. 11. Olsen BS, Sojbjerg JO, Dalstra M, Sneppen O. Kinematics of the lateral ligamentous constraints of the elbow joint. J Shoulder Elbow Surg. 1996;5:333-41. 12. Olsen BS, Sojbjerg JO, Nielsen KK, Vaesel MT, Dalstra M, Sneppen O. Posterolateral elbow joint instability: the basic kinematics. J Shoulder Elbow Surg. 1998;7:19-29. 13. Morrey BF, An KN. Functional anatomy of the ligaments of the elbow. Clin Orthop. 1985;201:84-90. 14. King GJW, Morrey BF, An KN. Stabilizers of the elbow. J Shoulder Elbow Surg. 1993;2:165-74. 15. Morrey BF. Surgical exposures of the elbow. In: Morrey BF, editor. The elbow and its disorders. 2nd ed. Philadelphia: WB Saunders; 1993. p 139-66. 16. Josefsson PO, Johnell O, Wendeberg B. Ligamentous injuries in dislocations of the elbow joint. Clin Orthop. 1987;221:221-5. 17. Milne AD, Chess DG, Johnson JA, King GJ. Accuracy of an electromagnetic tracking device: a study of the optimal operating range and metal interference. J Biomech. 1996;29:791-3. 18. King GJ, Zarzour ZD, Rath DA, Dunning CE, Patterson SD, Johnson JA. Metallic radial head arthroplasty improves valgus stability of the elbow. Clin Orthop. 1999;368:114-25. 19. Hannouche D, Begue T. Functional anatomy of the lateral collateral ligament complex of the elbow. Surg Radiol Anat. 1999;21:187-91. 20. Sojbjerg JO, Ovesen J, Gundorf CE. The stability of the elbow following excision of the radial head and transection of the annular ligament. An experimental study. Arch Orthop Trauma Surg. 1987;106:248-50.