Effects of pelvic obliquity and limb position on radiographic leg length discrepancy measurement: a Sawbones model

Purpose: Potential sources of inaccuracy in leg length discrepancy (LLD) measurements commonly arise due to postural malalignment during radiograph acquisition. Preoperative planning techniques for total hip arthroplasty (THA) are particularly susceptible to this inaccuracy, as they often rely solely on radiographic assessments. Owing to the extensive variety of pathologies that are associated with LLD, an understanding of the influence of malpositioning on LLD measurement is crucial. In the present study, we sought to characterize the effects of varying degrees of lateral pelvic obliquity (PO) and mediolateral limb movement in the coronal plane on LLD measurement error (ME). Methods: A 3-D sawbones model of the pelvis with bilateral femurs of equal-length was assembled. Anteroposterior pelvic radiographs were captured at various levels of PO: 0 degrees, 5 degrees, 10 degrees, and 15 degrees. At each level of PO, femurs were individually rotated medio-laterally to produce 0 degrees, 5 degrees, 10 degrees, and 15 degrees of abduction/adduction. LLD was measured radiographically at each position combination. For all cases of PO, the right-side of the pelvis was designated as the higher-side, and the left as the lower-side. Results: At 0 degrees PO, 71% of tested variations in femoral abduction/adduction resulted in LLD ME <0.5-cm, while 29% were >= 0.5-cm, but < 1-cm. ME increased progressively as one limb was further abducted while the contralateral limb was simultaneously further adducted. The highest ME occurred with one femur abducted 15 degrees and the other adducted 15 degrees. Similar magnitudes of ME were seen in 98% of tested femoral positions at 5 degrees of PO. The greatest ME (similar to 1 cm) occurred at the extremes of right-femur abduction and left-femur adduction. At 10 degrees of PO, a higher prevalence of cases exhibited LLD ME > 0.5-cm (39%) and >= 1-cm (8%). The greatest errors occurred at femoral positions similar to those seen at 5 degrees of PO. At 15 degrees of PO, half of tested variations in femoral position resulted in LLD ME > 1-cm, while 22% of cases produced errors > 1.5-cm. These clinically significant errors occurred at all tested variations of right-femur abduction, with the left-femur in either neutral position, abduction, or adduction. Conclusion: This study aids surgeons in understanding the magnitude of radiographic LLD ME produced by varying degrees of PO and femoral abduction/adduction. At a PO of <= 5 degrees, variations in femoral abduction/adduction of up to 15 degrees produce errors of marginal clinical significance. At PO of 10 degrees or 15 degrees, even small changes in mediolateral limb position led to clinically significant ME (> 1-cm). This study also highlights the importance of proper patient positioning during radiograph acquisition, demonstrating the need for surgeons to assess the quality of their radiographs before performing preoperative templating for THA, and accounting for PO (>5 degrees) when considering the validity of LLD measurements.