Background Measurement of joint space width (JSW) on antero-posterior (AP) radiograph of the pelvis is the reference imaging method for the diagnosis and follow-up of hip osteoarthritis (OA). However, the AP radiograph only profiles the apical roof of the acetabulum. Conversely, bilateral standing oblique (SO) radiographs of the pelvis can profile the full hip joint space provided that both hips are included in the field of view. Each SO view profiles the posterior horn, apical roof and anterior roof of the hip in contact with the radiographic table (“contact“ hip) and the apical roof, anterior roof and anterior horn of the contralateral hip, distant from the table (“distant“ hip). An important limitation to the use of the SO radiographs is the absence of reference measurements in a control population.
Objectives Aims of the present study were to define JSW normal values on the three hip views and to determine which among the different JSW measurement sites on the AP and the two SO views were the best discriminators between patients with clinical suspicion of hip OA and controls.
Methods JSW was measured at 10 points on the AP and bilateral SO views, in 37 hips of controls with back pain but no hip pain and 65 patients with a painful hip of less than one year duration.Patients were divided into two groups based on the absence (n=35) or presence (n=30) of degenerative changes of the subchondral bone. Intra-observer reproducibility was assessed. Descriptive statistics of average thicknesses at each pointon the three radiographs have been made in the three hips population. The ratio between JSW at the sites AR and PH for the contact hip on the SO radiograph was calculated. An ANOVA looked for a difference of the JSW between the three populations and between the populations two to two at each measurement site. A logistic regression evaluated how the JSW at each point predict the population of a hip. The cut-offthat maximizes accuracy was determined for each point where the ANOVA was significant and their properties were calculated.
Results NormalJSW measurements ranged from 5.02 to 4.28mm on the AP view, from 5.18 to 3mm on the SO view and from 5.39 to 3.83mm on the CSO view. The ANOVA showed significant difference in JSW between normal and painful hips without bony signs of OA only at two sites: PH on the SO view (p 0.02004) and IAH on the CSO view (p 0.00128). The p value (0.0687) was closed to the significance at the SAH site, on the CSO view. The JSW ratio AR/PH on the SO radiograph was also significantly different between the two groups (p 0.00446). No significant difference was found on AP X-ray. The logistic regression confirm good odds ratio for those points and especially for the AR/PH JSW ratio (0.0273), contrary to the JSW measurement on AP view. With a normal AR/PH ratio greater than 1, its sensibility and specificity were respectively 39 and 97%. At the other significant points, the sensibilities ranged from 35 to 75% and the specificities from 91 to 100%.
Conclusions JSW measurements on bilateral SO radiographs allow the diagnosis of early hip OA not shown by the standing AP pelvis radiograph. Measurement of the AR/PH JSW ratio provides the best diagnostic performances.
Disclosure of Interest None Declared