538 THE JOURNAL OF BONE AND JOINT SURGERY Upper limb Variation in the position and orientation of the elbow flexion axis A. Ericson, A. Arndt, A. Stark, P. Wretenberg, A. Lundberg From the Karolinska Institute, Stockholm, Sweden A. Ericson, MD, Consultant Orthopaedic Surgeon A. Stark, MD, PhD, Associate Professor P. Wretenberg, MD, PhD, Associate Professor Department of Orthopaedics, Karolinska Institute at Karolinska Hospital, 171 76 Stockholm, Sweden. A. Arndt, PhD, Head, Biomechanics Laboratory A. Lundberg, MD, PhD, Associate Professor Department of Orthopaedics, Karolinska Insitute at Huddinge University Hospital, 141 86 Stockholm, Sweden. Correspondence should be sent to Dr A. Ericson. ©2003 British Editorial Society of Bone and Joint Surgery doi:10.1302/0301-620X.85B4.13925 $2.00 e analysed the axis of movement in the normal elbow during flexion in vivo using radiostereometric analysis (RSA). The results show an intraindividual variation in the inclination of the axis ranging from 2.1˚ to 14.3˚ in the frontal and from 1.6˚ to 9.8˚ in the horizontal plane analysed at 30˚ increments. The inclination of the mean axis of rotation varied within a range of 12.7˚ in the frontal and 4.6˚ in the horizontal plane. In both planes, the mean axes were located close to a line joining the centres of the trochlea and capitellum. The intra- and interindividual variations of the axes of flexion of the elbow were greater than previously reported. These factors should be considered in the development of elbow prostheses. J Bone Joint Surg [Br] 2003;85-B:538-44. Received 14 October 2002; Accepted after revision 22 January 2003 The outcome of total joint replacement has improved over recent decades, particularly with regard to the hip and knee. 1-3 This can be attributed partly to improved bio- mechanical understanding of the joints and implants. Replacement of the elbow has become common in order to reduce pain and improve function in patients with rheuma- toid arthritis, but the results are proving to be less predicta- ble than those described for the hip and the knee. 4-8 The development of elbow prostheses has progressed from hinged to linked and unlinked designs. Critical analysis of the biomechanical features of the different implants is lack- W ing and there are no studies of the long-term results using a single design of prosthesis. Changes in design have there- fore been made without a thorough knowledge of the bio- mechanics. Dislocation associated with unlinked implants may be due either to incorrect positioning and orientation, or to defective design of the prosthesis. 9,10 The higher risk of loosening of linked implants may be a consequence of incorrect alignment of the axis of the hinge in relation to the anatomical axis of rotation. Furthermore, a semi-con- strained design usually allows about 5˚ of rotation between the components, which may be insufficient compared with the normal variation in the position and orientation of the joint axes. 11 Cadaver studies have shown that malposition- ing of either the ulnar or the humeral component modifies the kinematics of the artificial joint and may lead to loosen- ing. 9,10 In engineering terms, the movement of flexion and exten- sion of the elbow is referred to as a rotation of the olecranon round the trochlea of the lower end of the humerus. A pre- requisite for understanding the biomechanical properties of a joint is knowledge of the position of the axis of rotation. Early studies have described this as a stationary axis passing through the centre of the trochlea. 12-15 More recent studies, using electromagnetic tracking techniques, indicate that the axis is not fixed but changes in position and orientation throughout the arc of movement. 16-18 Most previous studies have been based on experiments in vitro. Their value, how- ever, is limited since movement may be dependent not only upon the geometry of the surface of the joint but also on physiological muscle function and on an intact capsule and ligaments. Knowledge of the position and orientation of the axis of movement in the normal elbow in vivo is indispensa- ble if improvements in the design of elbow prostheses and operative techniques are to be made. Radiostereometric analysis (RSA), so useful in the devel- opment of hip and knee replacements, 19-22 has been pre- dominantly applied to the measurements of movement between the implant and the host bone, in order to detect early loosening. It is also a highly accurate method for determining the axis of rotation in vivo even in a joint as complex as the ankle. 23 We undertook this study to deter- mine the position of the axis of rotation during flexion of the elbow in vivo in normal subjects.