Abstract

OBJECTIVES--To determine the feasibility of assessing patellofemoral osteoarthritis using the 'skyline' view and to compare its reproducibility with the standard lateral view. METHODS--Fifty patients attending a rheumatology outpatient department with osteoarthritis of the knee had standard radiographs taken of both knees: standing weightbearing anteroposterior; lateral supine radiograph in 30 degrees of flexion; and a skyline view of the patellofemoral joint. After an initial training period using 20 sets of films the remaining 30 sets were read blind by five observers. Intraobserver and interobserver variability was assessed using the kappa statistic. The minimum joint space in each compartment was measured using a ruler. Two views of a single normal subject were measured to determine the effect of knee flexion. RESULTS--The final 30 study films were from 20 women and 10 men, median age 72.5, range 18-91 years. A grading system comprising assessment of osteophytosis, joint space narrowing, sclerosis, cysts, and attrition could easily be applied to the skyline patellofemoral view. Intraobserver reproducibility was better than the interobserver reproducibility for all features. The skyline view performed more reproducibly and over a wider range of categorisation for joint space narrowing than the lateral radiograph. Measurement using a ruler was easy to perform and precise to within 1 mm for the medial tibiofemoral and lateral facet of the patellofemoral joint. In normal knees the degree of flexion significantly affected the measurement. CONCLUSIONS--Radiographic grading of the skyline patellofemoral view is readily achieved, is more reproducible than assessment of the lateral view, and allows more precise localisation of change. Such views should be considered in radiological surveys of osteoarthritis of the knee.