| dc.description.abstract | Following sustained kneeling in young adults, kinetic and kinematic changes at the knee and ankle during gait have been observed (Gaudreault et al., 2013; Kajaks & Costigan., 2015a; Tennant et al., 2014). It is possible that a change in the cyclic gait pattern after sustained deep knee bending (greater than 120 degrees of flexion) could deferentially load the cartilage in the knee and modify the mechanical stress exerted on the tissues, at least acutely, leading to increased risk of knee osteoarthritis development (Edd et al., 2018; Kajaks & Costigan, 2015). Gait changes that have been observed in those with knee osteoarthritis include decreases in knee flexion at heel strike, knee flexion range of motion, knee power, ankle power and knee flexion moment, and increases in knee adduction moment, and vertical loading rate. The primary purpose of this thesis was to explore whether differences in gait kinematics and kinetics following sustained kneeling mirror those occurring in osteoarthritic gait. It was hypothesized that prolonged kneeling would increase external peak knee adduction moment (KAM), and vertical loading rate (VLR), and would decrease knee flexion angle at heel strike (HS), knee flexion range of motion (ROM), external peak knee flexion moment (KFM) in early stance, positive peak ankle power (PAP) in late stance, and the second positive peak knee power (PKP) in late stance.
There is also a need to low-pass filter marker data to remove random noise, especially because noise is amplified when differentiating position and orientation for the calculation of linear and angular velocities and accelerations for inverse dynamics calculations such as joint moments and joint powers (Kristianslund et al., 2012; Sinclair et al., 2013). It has become common practice to filter raw marker and raw ground reaction force (GRF) data differently to preserve ground reaction force peaks and impact due to heel strike while walking. Papers reviewed for this project include a range of marker cut off frequencies from 5-8Hz and a range of GRF cut-off frequencies from raw (no filtering) to 40Hz. Selecting different cut-off frequencies for marker and GRF data can result in large oscillations in joint moments around heel strike during running that are not representative of the actual movement (Bezodis et al., 2013). It is possible that these effects could exist in any form of gait. The secondary purpose of this thesis was to determine the effects of various low-pass filter cut-offs on all dependent variables calculated for the primary objective. It was hypothesized that large differences in the filter cut-offs between marker and GRF data would result in significantly different knee flexion angles at HS, knee flexion ROM, first peak KAM, first peak KFM, PAP, PKP, and peak VLR.
As part of a previous study, fourteen participants (8M/6F) performed three pre- and post-kneeling gait trials at their self-selected pace while motion (Optotrak, NDI, Waterloo, Canada) and ground reaction force (AMTI, Watertown, MA, USA) data were recorded for the right leg. The kneeling protocol consisted of three ten-minute bouts of sustained plantarflexed kneeling separated by five-minutes of seated rest. Ground reaction force and motion capture data were filtered at 6Hz using a 4th order Butterworth filter using Visual 3D software (C-Motion Inc., Germantown, MD) to determine outcome measures. For the primary objective, seven two-tailed paired t-tests (one per outcome measure) were performed to compare subject mean values pre- and post- kneeling (α=0.05). For the secondary objective, marker and GRF data were filtered using a 4th order zero-lag Butterworth filter with 16 different cuff-off combinations. Each combination of GRF and marker input data was then used to calculate KFM, KAM, PAP and PKP for each subject’s pre- and post- kneeling gait trials. For measures where only kinematic data or only ground reaction forces were involved in their calculation (knee angle at HS knee ROM and peak VLR), only six different cut-off frequencies were used. A two-way repeated measures ANOVA was used to compare the mean outcome variables across different filtering conditions, and between the pre- and post-kneeling time points. Any main effect of filtering condition would indicate that filtering condition had a significant effect on the outcome measure. Any interaction would indicate that the effect of the kneeling exposure (part of the primary objective) depended on the filtering condition.
From the primary objective analysis, acute exposure to kneeling produced significant increases in knee flexion angle at HS, and peak KFM. An increase in knee flexion angle at HS and KFM peak (which typically occurs during early stance) suggests that participants could be attempting to reduce knee joint rate of loading more by accepting their weight with more knee flexion, as opposed to more knee extension which is typical of osteoarthritic gait and is related to greater axial loading rate at the knee. The increase in external KFM is balanced by the internal knee extensor muscle moment, suggesting that following kneeling, contradictory to what we hypothesized, participants are placing greater demand on their quadriceps. High flexion activities such as kneeling are associated with an increased risk of knee OA development. This work suggests that prolonged kneeling has the potential to compromise the integrity of the knee joint such that kneeling can acutely alter the loading patterns experienced at the knee joint (and thus potentially other lower limb joints) during the subsequent ambulation.
For the secondary objective of this thesis there was a significant main effect between pre- and post-kneeling for knee flexion at HS (p=0.034), and KFM (p=0.0063). The significant main effects between pre- and post-kneeling conditions for knee flexion at HS and KFM were expected, these results were found when investigating the primary objective. Interestingly there was also a main effect for filtering condition of peak VLR (p=0). The raw filtering condition was significantly larger than all other conditions, followed by the 25Hz being larger then 6Hz and 10Hz, and finally 20Hz being significantly larger then 6Hz. Since there was no significant main effect for filtering condition for any of the remaining dependent variables used, we can conclude that these peak measures are robust to filtering condition. With the lack of filter x kneeling condition interactions for all variables, we can infer that pre- and post- differences between subjects discreate variables are also robust to filtering condition. The lack of significant filtering condition effect for all outcome measures but the peak VLR is likely due to these measures being far enough away from heel strike, that the additional noise from oscillations caused by the impact at heel strike, like that seen during running, had minimal effect on these peaks except for peak VLR. This work shows that, in young asymptomatic participants, post-kneeling changes in kinetic and kinematic outcomes commonly used in the osteoarthritis literature are robust to changing filter cut-off frequencies. This finding may suggest that the wide variety of cut-off frequencies (when reported at all) in the literature on gait characteristics in osteoarthritis may not be of significant concern when comparing these outcome measures between studies. | en |
