Background The accurate tracking of knee joint motions during maneuvers associated with non-contact ACL injury is important for identifying injury mechanisms.

Objective We tested the hypothesis that motion capture and inertial measurement unit (IMU) measures of 3D changes in tibiofemoral angle and velocity are interchangeable in the ~70 ms weight acceptance phase of a 3–4*BW dynamic jump landing movement involving knee flexion and tibial internal rotation.

Design Two APDM IMUs and NDI Certus marker triads were rigidly attached to the mid-tibial and -femoral bone of cadaver knees to record motions during simulated jump landings. The initial knee angle was 15 degrees maintained by preparatory quadriceps muscle forces and tensile stiffness for the landing. The Bland-Altman Limits of Agreement (LoA) was used to compare the 3D data from 852 trials.

Setting University biomechanics research laboratory.

Patients Nine cadaveric knees harvested from six male and three female adult human donors.

Main Outcome Measurements The 3D knee angle changes from motion capture system were considered the gold standard and compared to calculated IMU data from the fusion algorithm provided by APDM Opal. The 3D tibial and femoral angular velocity changes measured by the IMUs were considered the gold standard and compared to the differentiated Certus angular data.

Results Although the mean peak IMU knee angle changes were slightly underestimated in all three orthogonal planes, the LoA bands were large, ranging from 35.9% to 49.8%. Certus had acceptable accuracy in the camera plane for angular velocity changes, with LoAs of ±54.9 ^o/sec and ±32.5 ^o/sec, respectively, for the tibia and femur.

Conclusions These IMUs could not reliably measure the peak 3D knee angle changes. Certus measurements of tibiofemoral angular velocity changes were comparable to IMU measures in the camera plane, and when velocities were sufficiently large.