Comparison of 5 Different Methods for Measuring Stress Radiographs to Improve Reproducibility During the Evaluation of Knee Instability

Background: Although stress radiography is frequently used to assess abnormal knee instability, the reliability and reproducibility for an evaluation of anterior-posterior instability of the knee may be affected by a variety of factors. Hypothesis: Different measurement methods result in different levels of reliability and reproducibility for instability; there may be a novel method that is more reliable and relatively unaffected by slight changes in flexion and rotation. Study Design: Cohort study (diagnosis); Level of evidence, 2. Methods: Stress radiographs of 40 patients with anterior cruciate ligament injury or reconstruction and 40 patients with posterior cruciate ligament injury or reconstruction were taken using the Telos device. The values of 4 conventional methods (medial-medial, lateral-lateral, middle-middle, and peripheral-central) and 1 newly devised method (Blumensaat line–anterior tibia) were compared. Intraclass correlation coefficients were examined to assess intraobserver and interobserver reliability of the measurements. For an evaluation of the reproducibility of each method, stress radiographs were taken twice (before and after the examination at the outpatient clinic) on the same day, and the values from the first and second stress radiographs were compared. Results: In the anterior drawer test, as to measurement reliability, the ranges of intraclass correlation coefficients were 0.713 to 0.889 for medial-medial, 0.624 to 0.812 for lateral-lateral, 0.834 to 0.932 for middle-middle, 0.722 to 0.892 for peripheral-central, and 0.891 to 0.963 for Blumensaat line–anterior tibia. As to test-retest reproducibility, the mean differences (SD) of displacement between the first and second radiograph were 1.0 (0.8) mm for medial-medial, 2.4 (2.3) mm for lateral-lateral, 1.7 (1.6) mm for middle-middle, 1.2 (0.6) mm for peripheral-central, and 0.5 (0.7) mm for Blumensaat line–anterior tibia. In the posterior drawer test, as to measurement reliability, the ranges of intraclass correlation coefficients were 0.859 to 0.958 for medial-medial, 0.773 to 0.915 for lateral-lateral, 0.859 to 0.951 for middle-middle, 0.852 to 0.958 for peripheral-central, and 0.893 to 0.961 for Blumensaat line–anterior tibia. Asto test-retest reproducibility, the mean differences (SD) of displacement between the first and second radiographs were 1.6 (1.3)mm for medial-medial, 1.8 (1.7) mm for lateral-lateral, 1.7 (1.5) mm for middle-middle, 1.4 (1.1) mm for peripheral-centr