BACKGROUND: Stress radiography is used in the valuation of soft tissue laxity following total knee arthroplasty (TKA). However, reliability and agreement is largely unknown.

METHODS: In this prospective reliability study, we included 15 participants with prior TKA. Standardized coronal stress radiographs were obtained in both extension and flexion and with both varus and valgus stress. All radiographs were repeated (test-retest). In extension the Telos stress device was used, and flexion radiographs were obtained using the epicondylar-view. Three independent raters measured angulation between femoral and tibial component from all radiographs. Reliability was assessed by intra-class correlation coefficient (ICC) and agreement visualized with Bland-Altman plots and by mean difference and limits of agreement (LOA).

RESULTS: Stress radiography in extension showed excellent reliability with ICC = 0.96 (0.95-0.98) and LOA of ±1.2°. Stress radiography at 80-90° of flexion showed good to excellent reliability when measuring medial laxity with ICC = 0.94 (0.89-0.97) and LOA of ±1.7°; however, when measuring lateral laxity the reliability was only moderate to good with ICC = 0.70 (0.51-0.84) and LOA of ±6.3°.

CONCLUSION: Stress radiography is clinically applicable and the methods described in this study provide excellent reliability for measurement of laxity in extension. The reliability of measurements in flexion is good to excellent when measuring medial laxity but only moderate to good when measuring lateral laxity.