Elbow kinematics : Studies of the elbow joint under normal conditions and after joint replacement

Sammanfattning: Background: Total elbow arthroplasty (TEA) is used for treatment of patients with severe pain and disability due to reumatoid arthritis. Long term results are not as good as for hip and knee replacements. Further development of the implant designs and surgical technique based on proper knowledge of the in vivo biomechanical joint properties are required to improve results. Aim: To analyse the variation and the position of the instantaneous flexion axes in vivo in the normal elbow joint and after TEA. Patients and Methods: Radiostereometric analysis (RSA) was used to determine the inclinations of the instantaneous elbow flexion axes in six healthy volunteers (study I) and in 13 patients with TEA (study III). Two of the implants were of a linked type (GSB III) and 11 were unlinked (five Capitellocondylar and six Kudo). Tantalum markers were implanted in the humerus and the ulna, defining two rigid bodies. Simultaneous radiographs were taken in maximum extension, 30, 60, 90, 120 and 150 degrees of flexion. The kinematic analysis defined the instantaneous flexion axes for each flexion increment and these were illustrated in standard drawings. Computed tomography (CT) was used in study II and IV to determine and visualise the position of the flexion axes relative to the individual joint in 3D. A spiral CT scan of the elbow was performed on the same subjects as in study I and on six of the patients that participated in study III (three Capitellocondylar and three Kudo prostheses). A linear algebraic solution was developed for a conform transformation between the RSA and CT data. Stability parameters were generated during the transformation. The calculated coordinates for the intersect points of the axes lateral and medial to the joint could be imported from RSA and designated in the 3D volume. The positions of the flexion axes could be visualised individually for each subject and patient by connecting the intersect points. Results: Study I. The range of the axis inclinations varied between subjects from 2.2º to 14.3º in the frontal and 1.6º to 9.8º in the horizontal planes. Mean axis inclination varied from 6,5º valgus to 6,2º varus, and from 2,4º internal to 2,2º external rotation. Study II. The median error between the transformed RSA coordinates and the CT coordinates was 0.22 mm, the rotation error 0.001º-0.006 ºand the scaling factor 0.985-1.035. The kinematic data for the instantaneous flexion axes were successfully transformed to the CT volume and axis position could be visualised in the 3D volume. Study III. The dispersion of axes varied for the unlinked prostheses from 4.1º to 84.3º (Kudo) and from 0.8º to 19.7º (Capitellocondylar) in the frontal plane. In the horizontal plane the dispersions varied from 3,3º to 45,0º and from 2,3º to 20,9º respectively. The two linked prosthesis (GSB) had axis dispersions of 13.0º and 15.4º in the frontal and 1,9º and 4,6º in the horizontal planes. Study IV. The prosthesis could be visualised in the CT volume with few or minor artefacts. All markers could be indentified and localised in the CT coordinate system. The RSA data could be fused with the CT data and the flexion axes visualised for the individual prosthesis. Conclusions: The use of RSA permitted determination of the inclination of the instantaneous elbow flexion axes in vivo in healthy volunteers and in patients after TEA. The proposed algorithm for fusion of RSA and CT data made it possible to also determine and visualise the 3D positions of the axes in the individual joint. As our proposed method for fusing RSA and CT data can be used in vivo in small series without further invasive technique, it can be of value for early in vivo assessment of new implants.