Weight-Bearing Asymmetry in Relation to Measures of Impairment and Functional Mobility for People With Knee Osteoarthritis

Participants

Participants for the OA group were recruited from the patients of three orthopedic surgeons at the University of Colorado Hospital prior to TKA from October 2008 to December 2009. Only patients with unilateral knee OA were included, defined by patients reporting no more than half the level pain on their non-affected knee compared with their affected knee (based on a numerical pain rating scale (NPRS) of 0-10). Volunteers were excluded if they were not within the age range of 50 to 85 years, had uncontrolled hypertension, uncontrolled diabetes, body mass index > 35 kg/m², neurologic impairment, had non-affected knee pain greater than 4/10, or other lower-extremity orthopedic problems.

Recruitment for the CTL group was made by verbal and written announcement in the community from December 2008 to March 2009. Exclusion criteria for CTL group volunteers were identical to the OA group with additional criteria that CTL volunteers be excluded if they did not exercise a minimum of 3 days per week (30 minutes/day), had knee pain greater than 2/10 on an intermittent basis (based on a NPRS of 0-10), or had any knee pain with regular daily activity.

The Colorado Multiple Institutional Review Board approved the study and written informed consent was obtained from all participants.

Weight-Bearing Asymmetry (WBA)

Measurement of WBA was achieved with vertical ground reaction force (vGRF) data obtained from two force platforms^a located under each foot of the participant during a timed Five Times Sit-to-Stand Test (FTSST)¹⁰ (see Functional Mobility Measures section for FTSST details). The force platform offsets were reset (i.e., zeroed) prior to each testing session. Average vGRF was calculated for each limb from start to end time of the FTSST. The vGRF values were normalized by dividing vGRF by body weight. To describe WBA in the OA group, the ratio of average vGRF value for the affected limb was divided by average vGRF for the non-affected limb to create a loading symmetry ratio (Equation 1).

Using a ratio equation to describe asymmetry is recommended when there is an identifiable affected side, as the results of this calculation are relatively easy to interpret.¹¹ For example, a ratio of 0.8 would indicate that the affected limb is loaded at a level 80% of the non-affected limb. For our study, the symmetry ratio allowed direct interpretation of WBA in terms of the relation between affected and non-affected limbs. However, for healthy individuals with no identifiable “affected” limb it would be arbitrary to select a limb for a ratio calculation. For example, a left over right limb ratio of 1.05, would be equivalent to a right over left limb ratio of 0.95. For this reason, we chose to use an absolute symmetry index (ASI)¹², which indicates absolute differences between limbs without regard to affected or non-affected limbs, to compare the OA and CTL groups (Equation 2). An ASI value of zero would indicate perfect symmetry.

Knee Impairment Measures

Three knee impairment measures were used in this study: self-reported knee pain, knee extensor strength asymmetry, and knee motion asymmetry. The first measure, self-reported knee pain, was assessed using a Numerical Pain Rating Scale (NPRS) from 0-10, with 0 representing no pain and 10 being the worst possible pain. Participants reported the level of pain for their affected knee immediately following the FTSST.

The second impairment measure was knee extensor strength asymmetry. Maximum isometric knee extension torque was measured using an electromechanical dynamometer^b. Participants were seated with their hips flexed at 85° and knees flexed at 60°. During this test, participants were instructed to provide their maximum knee extension effort and peak torque values were recorded. Two warm-up trials were performed prior to performance of 3 test trials, in which the torque values were recorded. For each limb, the peak torque value of the 3 test trials was used. Data from the electromechanical dynamometer were sampled at a frequency of 2000 Hz. During the measured trials, a monitor displayed a visual torque target that was at a slightly greater value than produced during the practice trials. Verbal encouragement was also used to elicit maximal effort. Knee extensor strength asymmetry was described using a symmetry ratio and ASI created from the peak torque values of each limb for the OA group and ASI for the control group.

The final knee impairment measured was joint motion, measured as maximal passive knee extension and flexion angles. Joint motion was measured using a universal manual goniometer with the participant positioned in supine. In this position, the proximal arm of the goniometer was aligned along the long axis of the thigh and the distal arm aligned along the long axis of the leg. To measure maximal passive knee extension, the participant's heel was placed on a block and pressure was applied by a tester on the distal thigh until a firm end feel was felt in knee extension. Maximal passive knee flexion was recorded as pressure was applied to the distal leg into knee flexion by a tester. Inter- and intra-tester reliability for goniometric measures of knee motion are high to very high (ICC range 0.70-0.99).^{13, 14} Maximal passive knee extension and flexion angles were used to calculate knee motion symmetry ratios and ASI values for the OA group and ASI values for the control group.

Functional Mobility Measures

Three standardized clinical measures were used to assess functional mobility: the Five Times Sit-to-Stand Test (FTSST), the Six-Minute Walk Test (6MWT), and the Stair Climb Test (SCT). The first measure, FTSST, was performed by each participant to quantify performance of transitions between sitting and standing.¹⁰ At the start of this test, participants were seated in a chair of standardized seat height (46 cm) with their feet placed on the center of each force platform in a participant-selected degree of comfortable knee flexion. The time (measured with a hand-held stop watch) taken to transfer 5 times between the sitting and standing positions as quickly as possible was recorded. The start and end point of the test was the participant seated with their back touching the backrest of the chair. Participants were given opportunity to practice the movement prior to testing and were instructed to perform the task as naturally as possible. Two trials were recorded and the fastest successful trial was used for analysis. Participants were encouraged to not use their hands during the test, neither on the chair arm rests nor the lower extremities. However, for participants who could not perform the movement without use of hands, armrest use was allowed and documented and the fastest time of the two trials was used for analysis.

The second functional mobility measure was the 6MWT, assessing gait function on a level surface. The 6MWT was originally developed as a measure of endurance for people with cardiovascular impairment.¹⁵ However, it is also a valid and reliable test for clinical measurement of gait function in a variety of populations, including individuals with knee OA.^16-18 Participants walked as far as possible during a 6 minute time frame. They were instructed to stop and rest if needed during the 6 minutes. The 6MWT was performed in a 30.5 m (100 ft) corridor.

The third functional measure, the SCT, assessed the time required to ascend and descend one flight of stairs (12 steps, 17.1 cm step height). Timing was performed with a hand-held stop watch and began when subjects lifted their foot for the first step and ended when both feet returned to the landing at the base of the stairs. The SCT has been shown to be a reliable and valid measure of the high level function required to perform stair ambulation for people with mobility limitations, including those with knee OA.^{19, 20} The shortest time required by each participant, out of two trials, was used.

Statistical Analysis

A sample size estimate for determining differences in WBA between the control (CTL) group and OA group was calculated on the basis of means and SD values available for the first 10 participants in the OA and CTL groups for the ASI. We determined the sample size for the OA group to be 50 based on the known number of 17 participants in the CTL group, a power level of 0.90, α level of 0.05, and a two-tailed t-test of difference between the group means.

Comparison of demographics and anthropometrics between the OA and CTL groups were assessed using independent t tests (for continuously scored variables) and a Fisher Exact Test (for sex distribution). An independent t test was also used to identify differences in WBA between the participants in the OA and CTL groups using the ASI. Knee extensor strength asymmetry and knee motion asymmetry were presented using the ASI for the OA and CTL groups, for visual comparison. Pearson product correlation coefficients were used to measure the bivariate relations between WBA and all other outcome measures. Unless otherwise specified, group values are reported as means (SD). The level of significance for the t test comparison of ASI between groups was set at p < 0.05. Confidence intervals were calculated for the interpretation of the Pearson's product correlation coefficients.

Group Comparison

Descriptive data for the OA and CTL groups are included in Table 1. The participants ranged in age from 51 to 85 years in the OA group and 58 to 81 years in the CTL group. Statistical comparison of sex, age, body mass, and body height revealed no significant differences between groups. Also included in Table 1 are the WBA, functional mobility measures, and knee impairment measures for both groups.

Weight-Bearing Asymmetry (WBA)

There was a significant difference (p value = 0.015) noted between the two groups in WBA during the FTSST (Table 1), with the OA group having ASI values indicating greater asymmetry than the CTL group. Figure 1 presents vGRF traces during the FTSST for a typical participant in the OA group. The loading symmetry ratio for the OA group was 0.87 (0.15), indicating a mean loading of the affected limb at 87% of the non-affected limb. Forty-two of the 50 participants (84%) in the OA group loaded less weight on the affected limb compared to the non-affected limb; 8 participants did not demonstrate affected limb unloading. Eighteen of the 50 participants in the OA group were unable to perform the FTSST without assistance of their upper extremities, and all but one of these 18 participants (94%) had lower average v-GRF values on the affected side. The average loading symmetry ratio for participants using their upper extremities for assistance was 0.77 (0.13) compared to 0.93 (0.14) for people not using their upper extremities.

Correlations of WBA with Outcome Measures

No correlation was found between symmetry ratios of maximum knee angle (extension or flexion) and WBA. Correlations of WBA with all other measures were fair²¹ (ranging from 0.29 to 0.44) and showed WBA to be positively associated with pain and knee extensor strength asymmetry and negatively associated with functional mobility. These associations are illustrated in the negative correlations of vGRF symmetry ratio (note: a low value of symmetry ratio indicates a high level of asymmetry) with knee pain, FTSST time, and SCT time, while positive correlations were identified for vGRF symmetry ratio with knee extensor strength symmetry ratio and 6MWT distance (Table 2).

Study Limitations

Causation for either the mechanisms related to WBA or the influence of WBA on function cannot be determined with correlations in a cross-sectional study. Due to potential confounding associations of pain, habitual movement patterns, impaired quadriceps femoris function, and impaired knee joint motion, it is important to separate these potential mechanisms for WBA before identifying which should be targets of clinical intervention. Future longitudinal studies designed to identify factors specific to knee OA are needed to further understand the mechanisms behind this functional problem and how it influences mobility. Additionally, our participants do not reflect the greater population of all individuals with knee OA. Participants in this study were all at the end-stage of unilateral knee OA, just prior to TKA surgery. A larger group of individuals across levels of OA severity is needed to examine WBA in relation to disease progression.

In order to promote a normal movement pattern between sitting and standing, participants were allowed to use their hands on the arm-rests if needed. However, it is possible that the 18 participants in this study who could not perform the FTSST without upper extremity support were compensating for the loading demands on the affected limb and that true WBA values were attenuated. Future investigation should examine if there is a maximum level of WBA that is achievable before upper extremity use is required.