Influence de la malposition rotatoire de l’implant condylien dans les échecs de prothèse unicompartimentale médiale du genou: Évaluation à l’aide d’un modèle géométrique dans une série de 276 interventions avec un récul de 7 à 15 ans

Abstract

Purpose of the study Rotatory malposition of the femoral component of a unicompartmental knee arthroplasty (UKA) is a key element of mechanical failure despite proper alignment and position of the tibial implant. The purpose of this study was to describe a method for measuring femoral implant rotation on the anteroposterior x-ray using an original geometric model. Material and methods 276 medial UKA (227 non-cemented Uni Goeland, Depuy; Uni, AMP) and 49 cemented (Miller-Galante, Zimmer) were reviewed and analyzed at a mean 11 years (range 7 - 15) using the Knee Society Score. Mean objective and functional scores were 43 and 47 points preoperatively. Postoperative alignment and position of the femorotibial contact point were determined. The frontal and sagittal position of the tibial plateau was noted by the angle formed with the mechanical axis. Angles α and β of the femoral implant formed between the mechanical axis and the greater condylar axis and the inferior condylar line passing through the apex of the condylar convexity respectively were measured on the AP view. Frontal rotation (angle β) of the femoral implant reflected the orientation of the distal condylar cut. The frontal measurement (angle α) was validated using simple geometric formulas: knowing angles α and β and the size of the implant and the displacement of the point of contact was measured together with the true rotation of the femoral implant or anteroposterior divergence, function of the difference α−β which is its frontal projection and reflects the orientation of the posterior condylar cut. Results Mean alignment was 3° varus. At last follow-up, outcome was satisfactory in 81.2% of the knees (224 UKA). Mean objective and functional scores were 90 and 87 points. Angles α and β were parallel and orthogonal to the mechanical axis or off by less than 4̊ external rotation (+6° to -4°). The mean difference α−β (frontal projection of the femoral implant transverse rotation) was 1°. The point of contact was 4 mm or less from the middle of the tibia. Failures were observed in 18.8% of knees (52 UKA): 4 for diverse reasons and 48 for mechanical failure, including 3 with polyethylene wear without loosening (two by eccentric point of contact and neutral position and one by inverted misalignments). 45 UKA (16.3%) presented loosening of the tibial plateau alone with rotatory femoral malposition: mean a angle off 13% from the mechanical axis (+16° to -9°), mean β angle off 8° (+12° to -8°), mean α−β difference 5°, femoral implant in square position on polyethylene with a peripheral point of contact ≥ 5 mm from the middle of the tibial plateau. The tibial implant exhibited mean varus of 1.5°. Discussion The main cause of failure was rotatory malposition of the condylar implant often associating frontal with transversal rotation. This increased mediolateral translation of the point of contact during knee motion causing abrasion and excessive pressure on the medial portion of the plateau. Frontal malrotation externalizes the posterior cut on the condyle tilted by the varus position, without correcting the varus of the mechanical axis in flexion. To avoid such prosthetic malrotations, three corrections must be made before making the bone cuts: cancel the external rotation of the condyles in flexion by positioning the distal cut guide in extension; re-establish the mechanical axis by reducing the internal tilt of the condyles resulting from the varus position both in extension (distal cut) and flexion (posterior cut). Improved instrumentations and reproducible techniques are needed to re-establish the 3D anatomic orientation of the femoral component, a key element for longer life of unicompartmental prostheses. This method for measuring the rotation of the femoral implant and the displacement of the point of contact using a geometric model applied to the AP view is useful for better understanding UKA failures.