Joint pain may cause patients to hold their limbs in mild flexion, abduction or adduction to minimize pain, regardless of the extent of articular pathology, and these positional changes may have substantial effects on the interpretation of radiographic joint space. We aimed to study the impacts of minor degrees of flexion, abduction or adduction of the hip, as well as the angle of the x-ray beam on the radiographic joint space width (JSW) of the hip joint. In the first part of the study, 65 patients (44 males, 21 females, mean+/-SD age 49+/-17) without clinical evidence of hip osteoarthritis (OA) who underwent intravenous pyelography (IVP) were studied. The hips were differentially positioned during the sequential radiographs required for the IVP procedure. Baseline radiographs were taken at 15 degrees internal rotation of the hips [the standard position for anteroposterior (AP) pelvis radiography]; additional positions included the hips at 15 degrees and/or 30 degrees flexion, and 15 degrees adduction and/or 30 degrees abduction. Radiographic JSWs were measured at three sites using a dial caliper: superomedial, superolateral, and the point of narrowest JSW. In the second part, 15 patients without clinical evidence of hip OA who underwent supine abdominal radiography for non-rheumatological indications were evaluated by standard (AP) pelvis x-ray in the same setting and JSW measurements were made as described above. When the average of the three measurements of the JSW was taken for each hip, baseline JSW was 4.38+/-0.55 mm (mean+/-SD). Positioning of the hip significantly (P<0.01) affected the radiographic JSW, with apparent widening during adduction and 30 degrees flexion, (JSW 4.56+/-0.51 mm and 4.53+/-0.58 mm, mean+/-SD), respectively, but narrowing during abduction (4.17+/-0.59 mm, mean+/-SD). Fifteen degree flexion of the hip did not result in statistically significant change in JSW measurements. Upon comparison of the AP pelvis view with the supine abdominal view, neither the average JSW nor the point of narrowest JSW differed significantly, although the superolateral JSW was significantly greater on the AP pelvis view (P=0.02). Subtle positional changes in the hip, such as may occur during pain or in OA, may artifactually alter the measured radiographic JSW. Thus, longitudinal studies which employ hip JSW to assess disease progression may yield biased results due to changes in pain rather than structure unless care is taken to ensure constant positioning of the hip.