Reliability and validity of knee angles and moments in patients with osteoarthritis using a treadmill-based gait analysis system

Abstract

BackgroundAlthough commonly used to study knee osteoarthritis (OA), relatively little is known about the reliability and validity of three-dimensional (3D) gait biomechanics derived from treadmill-based systems. Research questionUsing a treadmill-based gait analysis system, our objectives were to: 1) estimate the test-retest reliability of frontal and sagittal plane knee angles and moments in knee OA patients; 2) examine concurrent validity by estimating the associations between treadmill-based and overground (gold standard) measures; and 3) examine known-groups validity by comparing measures between knee OA patients and matched healthy controls. Methods34 patients and 16 controls completed 3D gait analyses using treadmill-based and overground systems. Treadmill walking speed was matched to self-selected overground speed. Marker set, knee angle and moment calculations were consistent for both systems. Patients completed a second test session using the treadmill-based system <24 h later but within 1 week of the first test session. Variables calculated from knee angle and moment gait waveforms during stance were evaluated using Bland and Altman plots, Intraclass Correlation Coefficients (ICC), Pearson correlations (r) and t-tests. ResultsVisual inspection of the Bland and Altman plots did not reveal any systematic differences between test and retest sessions; however, limits of agreement (LoA) were larger for the sagittal plane than the frontal plane. Mean differences between sessions for knee angles were <0.25 degrees and <0.18 %BW*ht for knee moments. ICCs ranged from 0.57-to-0.93 for test-retest reliability. Pearson correlations between treadmill and overground systems ranged from 0.56-to-0.97. Although highly associated, there were substantial differences in the moments, emphasizing they cannot be used interchangeably. Patients had greater first peak knee adduction moments (KAM) than controls [mean difference (95 %CI): 0.55 (−1.07, −0.04), p = 0.03]. SignificanceResults suggest frontal and sagittal plane knee angles and moments in patients with knee OA evaluated using a treadmill-based system are reliable and valid.