Jordy van Sambeeck

Age at surgery is correlated with pain scores following trochlear osteotomy in lateral patellar instability 5 73 Statistical analysis Descriptive statistics were used to summarize the data. Multivariable linear regression models were used to evaluate the association between age (independent variable) and VAS and KKS scores (dependent variables). Based on clinical experience and published literature we selected gender, history of surgery, presence of low or high grade trochlear dysplasia (low vs. high grade: A or C vs. B or D), postoperative patellar height, BMI and whether or not additional procedures were performed as independent variables that could influence patient reported outcome of surgery 26-28 . The multivariable regression analysis adjusts for these factors. Due to a lack of data, BMI as independent variable was not taken into account for analysis. A P-value of < 0.05 was considered statistically significant. All statistical analysis were performed using SPSS (v20, IBM SPSS Statistics, Armonk, NY, USA). Surgical technique A surgical technique was used as previously described by Koëter et al. 29 and slightly differs from the lateral facet elevating trochlear osteotomy as described by Albee and Weiker 30, 31 . In brief, the patient was placed supine on the table. Antibiotics were admitted preoperatively. No tourniquet was used. A lateral parapatellar incision was made and extended distally along the lateral femoral condyle. The retinaculum was opened in the direction of the femur. To visualize the osteotomy, two Kirschner wires were placed in the direction of the osteotomy till they were visible through the cartilage (halfway between themedial and lateral femoral facet). With the use of a small osteotome, an incomplete lateral trochlear osteotomy was carried out (figure 1). The curved osteotomy extended from the beginning of the trochlea proximally to the sulcus terminalis distally. Subsequently, the lateral articular surface of the trochlea was levered. In most cases, it was possible to raise the lateral articular surface by 4-6 millimeters. A wedge-shaped autograft was created with a part of the ipsilateral iliac crest to secure the elevation of the osteotomy; this graft was changed to a tricalcium phosphate (TCP) wedge during the study period (figure 2). Fixation of the osteotomy with osteosynthesis material was not needed. After performance of the osteotomy, the synovium was closed over while the lateral retinaculum was left open. Postoperatively, patients were placed on a continuous passive motion device (CPM) to stimulate a full passive range of motion until knee flexion was at least 60°. Patients were recommended the following training schedule: partial weight bearing for the first six weeks, without flexion limitation. After six weeks, full weight bearing was allowed. Patients were only referred to a physical therapist if restorage of normal gait was needed.