Measurement of leg length discrepancy after total hip arthroplasty. The reliability of a plain radiographic method compared to CT-scanogram

To measure the interobserver reliability and intraobserver reproducibility of post total hip arthroplasty (THA) leg length discrepancy (LLD) measurement on radiographs as well as to evaluate its accuracy by comparing it with LLD measurement on computed tomographic scanogram (CT-scanogram). In this prospective study, postoperative LLD measurements in ten THA patients were made by four observers on anteroposterior radiographs of the pelvis (inter-teardrop line to the tip of lesser trochanter) and compared to LLD measurements made on CT-scanogram scout views of the lower limb. Two observers repeated the LLD measurements on radiographs 8 weeks after the first measurements. The interobserver reliability of the LLD measurement on plain radiographs was evaluated by comparing the measurements of the four observers and the intraobserver reproducibility by comparing the two repeated measurements made by the two observers. We found excellent interobserver reliability (mean ICC 0.83) and intraobserver reproducibility (ICC 0.90 and 0.88) of the LLD measurements on plain radiographs. There was a moderate to excellent agreement, but with wide variation of measurements among the four observers, when plain radiographic measurement was compared with CT-scanogram (ICC 0.58, 0.60, 0.71, and 0.82). Despite the excellent interobserver reliability and intraobserver reproducibility of LLD measurement on radiographs, clinicians should be aware of its limited accuracy when compared to CT-scanogram. (orig.)