SEARCH

SEARCH BY CITATION

Keywords:

  • Force steadiness;
  • Force accuracy;
  • Proprioception;
  • Eccentric strength;
  • Pain

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Objective

To characterize the distribution of error in knee joint proprioception, quadriceps force accuracy and steadiness, and muscle strength in patients with knee osteoarthritis (OA). Special attention was paid to eccentric strength.

Methods

We compared knee OA patients (n = 20: 15 women, 5 men) with age- and sex-matched, symptom-free adults. Knee pain and mobility were assessed with standard tests. Knee joint proprioception was measured with a repositioning test. Quadriceps force accuracy and steadiness were determined during a force target-tracking task. Maximal voluntary quadriceps force was measured during eccentric, isometric, and concentric contractions.

Results

OA patients had knee pain, needed 67% more time to complete 4 functional tasks, and produced 82% more proprioception errors (all P < 0.05). About 80% of this error was due to overshooting the target and 68% of the overshooting error occurred at 2 of the 5 least flexed knee joint positions. OA patients had 89% more errors in accurately matching target forces during submaximal quadriceps contractions and in the same tasks, OA patients also produced these forces with 155% more variability (all P < 0.05). OA patients had especially weakened ability to produce maximal voluntary eccentric strength.

Conclusion

Quadriceps dysfunction in knee OA includes impaired proprioception, especially in the more extended knee joint positions; impaired ability to accurately and steadily control submaximal force; and impaired eccentric strength. These results have implications for designing exercise and rehabilitation programs for patients with knee OA.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Osteoarthritis (OA) of the knee is associated with disability due to pain, quadriceps dysfunction, and impaired proprioception. OA patients have up to 4° of error in their ability to reposition their limb and to perceive movement (1–11). However, little is known about the distribution of this error in the range of motion of the knee. Such information would be clinically important because from the distribution of the error one could infer the type of malfunctioning receptors and improve the design of rehabilitation protocols. Joint receptors contribute to position sense primarily at the extremes of the joint range of motion and muscle receptors discharge to indicate limb position in the midrange; therefore, it is important to determine where in the range of motion the proprioceptive error is the largest (12, 13). Thus, one aim of the present study was to determine the distribution of error in knee joint proprioception in age- and sex-matched subjects with and without knee OA.

Injury to a joint affects the magnitude of maximal strength as well as the control of force produced by the muscles surrounding the joint (14, 15). Adequate control of submaximal muscle forces is especially important in activities of daily living (ADLs) that are normally executed at a fraction of the available maximal muscle strength in young and middle-aged adults (16). Accuracy and steadiness of force production are 2 forms of force control that have been used widely to assess the quality of force production in geriatric populations in a variety of muscles and tasks (17–19). However, it is unknown if the quantitative reduction of maximal strength of the quadriceps muscle in patients with knee OA is accompanied by a concomitant impairment of submaximal forces. Such information is clinically relevant for understanding locomotor deficits and improving currently used exercise rehabilitation programs.

Another hallmark symptom of knee OA is quadriceps weakness. Compared with age- and sex-matched healthy adults, patients with knee OA present with 10–60% less quadriceps muscle strength (8, 20–22). Quadriceps weakness is normally quantified as a reduction in maximal isometric and isokinetic concentric muscle strength. However, ADLs also comprise eccentric or lengthening contractions and it is unclear if quadriceps weakness is present to the same extent in this expression of maximal eccentric strength (23). This is an especially relevant issue for knee OA patients who have quadriceps weakness due partly to incomplete muscle activation (8, 24); even healthy young adults are often unable to fully activate their muscle during a maximal effort eccentric contraction (25). In addition, in eccentric contractions, some researchers observed a preferential recruitment of type II muscle fibers (26) and a greater selective atrophy of type II muscle fibers in muscles around joints with OA compared with muscles of age-matched controls (27, 28). Thus, it is reasonable to hypothesize that patients with knee OA would exhibit weakness in eccentric strength.

Taken together, the aim of the present study is to compare the distribution of error in knee joint proprioception, force accuracy and steadiness, and the force-velocity relationship in patients with grade II knee OA and age and sex-matched healthy adults. Special attention will be paid to eccentric strength. The characterization of these aberrations represents a shift from the traditional quantification of quadriceps dysfunction to the identification of novel control properties of quadriceps force production in patients with knee OA.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Subjects.

Table 1 shows the characteristics of the 2 subject groups (n = 40) recruited through advertisements in local newspapers. After a telephone interview, potential subjects were interviewed in person and confirmed that the participant, except for knee OA, was ostensibly healthy and free from known neuromusculoskeletal illnesses or injuries. All knee OA subjects had their diagnosis confirmed by a physician and had Kelgren and Lawrence grade II or higher bilateral knee OA in the tibiofemoral compartment (29). OA subjects exhibited chronic and stable knee pain; radiographic signs of hypertrophic changes, marginal spur formation, subchondral sclerosis or cyst formation, and nonuniform joint space narrowing; and had difficulty rising from a chair and ascending or descending stairs. Exclusion criteria were heart disease, stroke, insulin-dependent diabetes, osteoporosis, rheumatoid arthritis, using medication that caused dizziness, unstable medication schedule, history of falls or other motor deficits, severe recent modifications of diet, inability to walk up and down a flight of stairs, corticosteroid injection within the past 30 days, inability to comprehend and follow instructions, and an apparent lack of commitment to participation.

Table 1. Subject characteristics*
VariableOsteoarthritis (n = 15F, 5M)Controls (n = 15F, 5M)
  • *

    Data presented as mean ± SD (range). BMI = body mass index.

Age, years57.5 ± 7.3 (43–68)56.8 ± 5.0 (45–66)
Mass, kg86.8 ± 18.3 (62–110)81.2 ± 12.5 (60–96)
Height, meters1.64 ± 0.09 (1.50–1.82)1.67 ± 0.07 (1.52–1.83)
BMI, kg/m229.3 ± 3.2 (24–35)28.3 ± 3.0 (23–33)

Table 1 also shows the characteristics of the healthy comparison group. Except for knee OA, knee pain, and poorer mobility, the subjects in this comparison group met the same exclusion and inclusion criteria as the OA patients. All subjects provided written informed consent in accordance with University policy and the University Institutional Review Board approved all research procedures.

Testing procedures.

General testing protocol.

On their first visit to the laboratory, subjects were interviewed, signed the consent form, and were familiarized with the testing equipment and environment. Subjects were then tested in a single, 1.5-hour session. However, 12 subjects from each group were tested twice, the 2 sessions separated by a period of 10 weeks, to assess reliability of the dependent variables. On the day of testing, subjects' height and weight were measured and their level of knee pain recorded. As a general warmup, each subject rode a bicycle ergometer at 60 revolutions per minute for 5 minutes at 0.5–2.0 kg resistance and performed 3 minutes of lower extremity stretching. Next, functional mobility, knee joint proprioception, quadriceps force accuracy and steadiness, and maximal quadriceps strength were measured. Measurements were performed on the leg for which the subject reported more pain (OA group) or on the dominant leg (control group), determined by kicking a ball. There were about 5 minutes of seated rest between tests.

Knee pain.

Long-term history of knee pain was established during the interview. Subjects rated their knee pain in each leg 1 week before testing and immediately before and after the testing session using a 5-point scale (0 = no pain, 1 = slightly painful, 2 = moderately painful, 3 = very painful, 4 = excruciatingly painful) during level walking, walking up stairs, walking down stairs, and rising from a chair. Three dependent variables of pain were computed; mean pain scores for the week prior to testing and for the day of testing immediately before and after the session.

Functional mobility.

Subjects walked 18 meters on a carpeted and level floor of an indoor hallway. They walked down and, after a minute of rest, up a 12-step stairway at a safe, self-selected pace. Lighting was adequate and subjects could use the banister rails. The fourth mobility task was the get-up-and-go test timed with a digital stopwatch (30).

Knee joint proprioception.

Proprioception was assessed in a joint repositioning task. Subjects sat on the dynamometer's seat (Kin-Com, AP125; Chattecx Inc, Chattanooga, TN) with knee and hip joints at 90°. Crossover shoulder straps, a lap belt, a knee strap, and an ankle cuff minimized extraneous movements. The lever arm pad was strapped around the lower leg, just above the malleoli. The dynamometer's axis of rotation was aligned with the knee joint's axis of rotation. The dynamometer was programmed so that subjects were able to move their leg, attached to the lever arm, without resistance from the dynamometer. The dynamometer's electrogoniometer was used to measure knee joint position to 0.5° accuracy and the signal was sampled at 1 kHz. Subjects completed 2 trials at each of the 5 positions, resulting in 10 total trials.

Every trial started from 90° of knee flexion. The investigator passively moved the subject's leg, strapped to the lever arm, by grasping the lever arm attachment and moving it to 1 of 5 target positions. The investigator held the limb for 5 seconds, reminded the subject to “remember” the specific knee joint position, and moved the leg back to the 90° starting position. Subjects practiced the repositioning task 5–10 times with eyes open. Then the subjects were blindfolded for the testing. At random order, proprioception target positions were set at 15°, 30°, 45°, 60°, and 75° of knee flexion. The dependent variable was computed as the difference between the actual leg position and the target position and it was expressed as the average absolute error in degrees.

The direction of joint proprioception error was also determined for each of the 5 joint positions in each subject (2 trials at each of the 5 positions, 10 trials per subject). An overshooting error represented a more extended knee position than the criterion position. An undershooting error represented a more flexed knee position than the target position. A null error represented that the subject was on target. The total number of responses was then used in the chi-square analysis.

Quadriceps force accuracy and steadiness.

Except for the target forces being 50N and 100N in the present study, we previously provided a detailed description of subject preparation, subjects positioning, number of repetitions, data analysis, and reliability of the force accuracy and force steadiness measurements (19). In brief, force variability was expressed as the standard deviation in Newtons and force accuracy as the mean absolute error relative to the target force during 5-second eccentric, isometric, and concentric quadriceps contractions. Figure 1 shows representative force-time trials from 1 individual with and 1 individual without knee OA.

thumbnail image

Figure 1. A representative trial of an eccentric contraction in a 49-year-old woman with knee osteoarthritis (OA; thick line) and a 50-year-old female control subject (thin line). The horizontal line represents the 50-N target force. The difference between the force signal and the target line represents the absolute force error, or the measure of force accuracy. The fluctuation in the force signal, expressed as the standard deviation, represents force steadiness. In these trials, the average absolute error was 24N and the force steadiness, expressed as the standard deviation of force, was 16N in the knee OA subject. In the healthy subject, the corresponding values were 10N and 9N. s = seconds.

Download figure to PowerPoint

Maximal voluntary strength.

After the accuracy and steadiness tests, maximal voluntary isometric and isokinetic eccentric and concentric (90°/second and 180°/second) quadriceps strength were measured on the same dynamometer described above. We previously provided a detailed description of subject preparation, subject positioning, number of repetitions, data analysis, and reliability of these tests (19).

Statistical analyses.

Reliability of the dependent variables was estimated with a paired t-test and intraclass correlation coefficient in subsets of 12 healthy adults and 12 OA subjects, respectively. The key analyses included a comparison between the OA and the control group using an analysis of variance (ANOVA) with repeated measures on each dependent variable, respectively, for proprioception, force accuracy, steadiness, and muscle strength. Significant F values from the ANOVAs were analyzed with a Tukey's post-hoc contrast to determine the means that were different at P < 0.05. The direction of joint position error was analyzed with a 2-group (OA, healthy) by 3 types of response (overshooting, undershooting, on target) chi-square analysis. The relationship between 2 variables was determined with a Pearson's product-moment correlation coefficient or linear regression.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Reliability.

Table 2 shows that the reliability of the dependent variables was acceptable based on the test-retest data from the healthy adults and from knee OA subjects.

Table 2. Test-retest reliability of the dependent variables*
VariableOsteoarthritis (n = 12)Healthy (n = 12)
Change, %RChange, %R
  • *

    Values are percent change between test 1 and test 2 and the associated intraclass correlation coefficient. Functional tasks include 18-meter walking, ascending and descending 12 steps, and get-up-and-go test. MVC = maximal voluntary contraction.

Functional tasks−5.00.95−1.00.79
Proprioception11.10.72−7.30.79
Knee pain4.90.93--
MVC    
 Eccentric−1.80.94−5.00.87
 Isometric0.00.95−9.20.88
 Concentric7.30.822.60.97
Force accuracy    
 Eccentric−6.40.852.30.91
 Isometric3.30.874.60.78
 Concentric7.70.92−3.80.84
Steadiness    
 Eccentric5.90.88−5.60.82
 Isometric1.30.94−1.10.93
 Concentric−2.80.846.30.85

Pain, functional mobility, and proprioception.

All subjects completed the testing session without complications. On a scale of 0–4, OA patients reported similar levels of pain for the week before testing (mean ± SD: 1.87 ± 0.32), immediately before (1.96 ± 0.38), and immediately after (1.86 ± 0.29) the testing session (one-way ANOVA, F = 1.1, P = 0.3482). Healthy adults reported no knee pain. On average, OA subjects needed 67% more time to complete the functional tasks (Table 3).

Table 3. Group data for time to complete functional tasks*
TaskOsteoarthritis (n = 20)Healthy (n = 20)Difference
AbsolutePercentage
  • *

    Values are group means ± SD and are expressed in seconds.

  • P = 0.0001 for unpaired t-test between the 2 groups.

Level walking17.6 ± 2.111.5 ± 1.16.152
Stair descent10.2 ± 2.35.8 ± 0.94.476
Stair ascent12.8 ± 3.27.2 ± 0.85.678
Get-up-and-go9.3 ± 1.15.7 ± 0.73.663
Mean12.5 ± 2.27.6 ± 0.95.067

Table 4 shows the significant group by joint position interaction for average absolute error as a measure of knee joint proprioception. Compared with healthy controls, OA patients produced more error when the knee was in a less flexed position (F = 12.4, P = 0.0014). Overall, OA subjects had 2.4° greater average absolute error to reposition their leg to the 5 targets. In addition, of the 200 repositioning attempts by the OA group, 157 (78.5%) resulted in overshooting, 28 (14%) in undershooting, and 15 (7.5%) attempts were on target. Of the 200 attempts in healthy subjects, 74 (37%) resulted in overshooting, 20 (10%) in undershooting, and 106 (53%) were on target (χ2 = 99.6, P < 0.0001). OA subjects produced 68% of the overshooting error at target knee joint positions of 15° and 30°.

Table 4. Group data for knee joint proprioception*
Knee flexion, degreesOsteoarthritis (n = 20)Healthy (n = 20)Difference
AbsolutePercentage
  • *

    Values are group means ± SD and are expressed as absolute average error in degrees. Significant group by joint position interaction (F = 12.4, P = 0.0014).

  • P < 0.05 compared with absolute average error at the other 3 joint positions based on Tukey's post-hoc contrast.

  • P < 0.05 between groups based on Tukey's post-hoc contrast.

156.7 ± 2.83.1 ± 1.33.6117
306.0 ± 2.22.8 ± 1.33.2114
454.8 ± 1.63.2 ± 1.31.650
604.6 ± 1.52.5 ± 0.92.182
754.1 ± 1.12.8 ± 1.21.347
Mean5.3 ± 1.92.9 ± 1.22.482

Quadriceps force accuracy and steadiness.

The upper half of Table 5 shows the results for the group by contraction type by force level 3-way interaction (F = 5.9, P = 0.0207). Compared with control subjects, OA patients produced 136% more error during eccentric (P < 0.05) and 107% more error during concentric contractions (P < 0.05). Overall, OA patients had 89% greater average absolute error than control subjects. Subjects in both groups produced isometric force accurately.

Table 5. Group data for quadriceps force accuracy and steadiness*
Muscle contractionTarget force, NOsteoarthritis (n = 20)Healthy (n = 20)Difference
AbsolutePercentage
  • *

    Values are group means ± SD. Accuracy of force production is expressed as the average absolute error (in N) relative to the target force. Group by contraction type by force level 3-way interaction (F = 5.9, P = 0.0207). Force steadiness is expressed as the standard deviation of the absolute force fluctuation (in N).

  • P < 0.05 compared with isometric and concentric contraction based on Tukey's post-hoc contrast.

  • P < 0.05 between groups based on Tukey's post-hoc contrast.

  • §

    P < 0.05 compared with isometric contraction based on Tukey's post-hoc contrast.

Accuracy     
 Eccentric5026 ± 1010 ± 416160
 10017 ± 88 ± 39§112
 Isometric503 ± 12 ± 1150
 1003 ± 13 ± 100
 Concentric5017 ± 48 ± 29§113
 10012 ± 26 ± 18§100
Steadiness     
 Eccentric5015 ± 57 ± 38114
 10016 ± 34 ± 212300
 Isometric504 ± 13 ± 1133
 1004 ± 13 ± 1133
 Concentric508 ± 3§4 ± 24§100
 1009 ± 2§2 ± 17§350

The lower half of Table 5 shows the group by contraction type by force level interaction for force steadiness (F = 7.3, P = 0.0111), i.e., subjects were least steady, in order, during eccentric, concentric, and isometric contraction. Overall, the OA patients were 155% less steady than controls. Force production was especially unsteady during eccentric (15.5N) compared with concentric contraction (8.5N; P < 0.05). OA patients were unsteady at both 50-N and 100-N target forces during dynamic contractions, whereas control subjects demonstrated poorer steadiness at the lower target force. Both subject groups exhibited similarly steady isometric force.

Maximal voluntary quadriceps strength.

Figure 2 shows the group data for the force-velocity relationship and the group data by condition interaction (F = 44.7, P = 0.0001). OA patients produced 76% lower eccentric forces compared with the 56% lower forces under isometric and concentric conditions. OA patients' eccentric weakness was 20% greater than the weakness in concentric strength. Overall, OA patients produced 63% less quadriceps force than control subjects (P < 0.05).

thumbnail image

Figure 2. Group data for maximal isokinetic quadriceps force in subjects with knee osteoarthritis (OA; filled symbols) and age- and sex-matched healthy adults (open symbols). Note the flattened force-velocity relationship in the OA group, suggesting an impaired ability to produce eccentric force. *P < 0.05 between groups. †P < 0.05 compared with eccentric force in healthy subjects. ‡P < 0.05 compared with isometric and eccentric force in OA patients. N = Newtons; s = seconds. Error bars show SD.

Download figure to PowerPoint

Table 6 shows the results of the correlation analyses. The correlation coefficients ranged from r = –0.68 (between muscle strength and time to execute functional tasks) to r = 0.62 (between functional time and force unsteadiness).

Table 6. Correlations among the dependent variables*
 KPFTPEMSFEFU
  • *

    KP = knee pain in arbitrary units; FT = functional time, expressed as the summed time to execute the 4 functional tasks; PE = proprioception error, expressed as sum of the errors in repositioning the knee joint to 5 targets; MS = maximal voluntary quadriceps strength, expressed as sum of eccentric, isometric, and concentric force; FE = force accuracy error, expressed as the sum of absolute errors during isometric, concentric, and eccentric contractions at 50-N and 100-N target forces; and FU = force unsteadiness, sum of the standard deviation scores of force steadiness at the 2 force levels and 3 contraction types. Knee pain only in OA patients (n = 20), otherwise n = 40.

  • P < 0.05.

  • P < 0.01.

Knee pain-0.53−0.01−0.210.080.11
Functional time -0.57−0.680.330.62
Proprioception error  -−0.210.410.39
Maximal strength   -0.00−0.13
Force accuracy error    -0.40
Force unsteadiness     -

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The finding of 2.4° impairment of proprioception in patients with knee OA is consistent with the 1–4° error found using similar (2, 9) or other methods of assessing proprioception (2, 3, 10, 11). A new finding in the present study is that subjects with knee OA produce more joint position error at 2 of the 5 positions in which the knee is in the most extended positions. In contrast, healthy subjects have a similar magnitude of error at different knee positions throughout the range of motion. In addition, nearly 70% of the error in OA patients came from overshooting the target. So far it is unknown if the magnitude of knee joint proprioception error is evenly distributed in the range of motion. Previous studies determined the error at specific knee joint positions and reported the average error across the entire range of motion (2, 9–11). In a viscosupplementation study, knee OA patients, as in the present study, produced significantly more error in more extended knee joint positions (31). However, that study did not use a healthy control group and the OA patients were older and more debilitated than our OA patients.

The findings that proprioception errors occurred in the more extended knee positions and were due to overshooting the target indicate a reduction in the sensitivity of the proprioreceptors in OA subjects' knee joints. These receptors could include the quick-adapting Pacinian corpuscles, which sense joint motion, and the slow-adapting Ruffini receptors, which sense joint position. In the more extended knee positions, the quadriceps muscle is at a short length and the source of the proprioceptive dysfunction would be less likely to emanate from the Ia afferent muscle receptors, which primarily contribute to proprioception at longer lengths of muscle (11–13). Another possibility is that in the extended knee positions, reciprocal inhibition from the quadriceps to stretched hamstrings is not as effective in knee OA, contributing to the proprioception errors (32, 33).

This study is the first to demonstrate a link between knee proprioception error in knee OA and performance in ADLs (3). We observed a moderate relationship of r = 0.57 (P < 0.01; Table 6) between the average proprioception error at the 2 most extended knee positions and the time to execute ADLs, without a significant association between knee pain and proprioception errors. OA patients must select a configuration of joint positions for gait that produces the least discomfort. Slow gait facilitates this selection. The argument was made that an error of only 1–1.5° in knee joint position at heel strike and the late phase of swing during gait could contribute to OA subjects' altered gait and to the exacerbation of knee OA (8, 34). Although knee joint position at heel strike during gait may not differ between OA and healthy individuals (35), the amount of maximal flexion is smaller by 3–4°, resulting in a more extended limb during stance (36, 37). In addition, OA subjects compared with healthy adults had up to 2-fold greater variability in knee joint flexion while executing ADLs (35). Impaired proprioception can contribute to this increased variability. Taken together, OA patients have an aberrant proprioception system that increases the potential for positional errors in ADLs and reduces the potential to sense and correct such errors. Proprioception errors manifest in ADLs by increasing the variability around the correct limb position and contributing to altered gait (8, 34).

Force accuracy and steadiness describe the quality of submaximal force production. There is strong evidence to suggest that muscle force accuracy and steadiness decline with age (17–19), but it is unknown if such an impairment would also occur in knee OA. Our expectation that OA subjects do exhibit such impairment is based on the observation that individuals with a joint malfunction or injury present with some form of impaired force control (8, 11, 14, 24, 38). Indeed, our data indicate that OA patients had 89% more error in a force-matching task and had 155% greater variation in producing force smoothly. These data document a novel qualitative aberration in the quadriceps muscle of knee OA subjects and expand on the findings of many studies that reported quantitatively reduced maximal voluntary strength in knee OA (8, 20–23, 38–40).

The error in accurately matching the 50-N and 100-N target forces resulted from overshooting the target. Force accuracy generally improves with increasing force levels but when it is present, the error is almost always in the positive direction (17, 19). The OA group also followed this pattern and exerted more force than was necessary to reach the target under dynamic conditions.

We also observed significantly less steady force production of the quadriceps muscles in patients with knee OA compared with healthy subjects. Force steadiness was similar at the 2 target forces of 50N and 100N, even after the steadiness data were normalized for force and expressed as coefficient of variation (data not shown). By coincidence, the 50-N target force in the OA patients and the 100-N target force in healthy subjects represented about the same relative force of the maximal eccentric (∼17%) and concentric (∼24%) voluntary contraction, and the large differences in steadiness still persisted at the same relative force levels. These data suggest that the impaired ability of the quadriceps muscle to control force in OA patients is not the result of OA patients' reduced maximal quadriceps strength.

In previous reports, force steadiness improved with increasing force levels in old adults (for review see reference 18), but we found that OA patients' force steadiness was equally impaired at low and moderate force levels. Thus, the possibility exists that OA patients have difficulty in producing force smoothly with their quadriceps muscle over a broader range of forces than healthy age-matched adults (Table 5) and old adults (18, 19). Compared with healthy subjects, force steadiness in OA patients was especially impaired during eccentric contraction, less impaired during concentric, and unimpaired during isometric contraction. Coupled with their reduced ability to produce maximal eccentric force, the poor eccentric force steadiness and accuracy in OA patients suggest a generalized impairment of generating and controlling maximal and submaximal eccentric quadriceps force. This conclusion is strengthened by the observation that the variability in OA patients' joint torques in ADLs was up to 3.9-fold greater (35).

There is extensive documentation in the literature that individuals with knee OA have a large deficit in isometric and concentric quadriceps muscle strength (8, 20, 21, 24). The results of the present study expand on these findings and suggest that the ability of OA patients to produce force while the quadriceps muscle lengthens is significantly more impaired than force production under isometric and concentric conditions.

The eccentric weakness in knee OA is probably of neural origin. Slemenda et al reported that quadriceps weakness was independent of muscle size in knee OA (21), and others reported that OA subjects were unable to fully activate the quadriceps during maximal isometric contractions (8, 20). Indeed, healthy individuals also have difficulty in fully activating their muscles during a maximal voluntary eccentric contraction due to neural inhibition associated with the high forces (25, 41). It is likely that eccentric weakness in knee OA is the combined effect of the natural inhibition associated with an eccentric contraction and the arthrogenous muscle inhibition uniquely present in knee OA. A parallel mechanism could be related to motor unit recruitment. In eccentric contractions, a preferential recruitment of type II muscle fibers occurs (26). Even though total quadriceps muscle size may be similar between individuals with and without knee OA (21), selective atrophy of type II muscle fibers has been observed in muscles around joints with OA (27, 28). This could partly account for the low eccentric strength in patients with knee OA.

The present data suggest that clinicians should evaluate proprioception in knee OA at multiple points in the range of motion, with special attention to errors in the more extended knee joint positions. Reduced knee joint proprioception in OA patients indicates the need to incorporate proprioception training in OA patients' exercise program to slow disease progression. It is unknown whether exercise can improve joint proprioception in patients with knee OA. Proprioception training of the ankle or knee after a ligament injury produced mixed results, probably due to the low specificity of the proprioception training in relation to the high loading rates of the joints used in the target tasks (42–45). Conceptually, OA patients are ideal candidates for training that uses a combination of weight-bearing and non–weight-bearing tasks for proprioception training because they are free of ligamentous injury and the loading rate of the knee joint during their slowed gait is also low (42). Although not a causative relationship, the correlation between proprioception errors and function suggests that improved proprioception may lead to improved function by improving the awareness of limb position in space, reducing pain, and increasing muscle strength (15, 42).

There was a moderate (r = 0.62) correlation between the time to execute ADLs and variability of force (steadiness), suggesting clinical relevance of impaired force control for daily function. Accurate scaling of eccentric force of the quadriceps is vital for successful execution of ADLs and it affects our ability to achieve the necessary force and alters the kinematics of the limb. For example, stair descent is dominated by eccentric quadriceps contractions (46), and this ADL is associated with the highest rate of falling (47). In addition, the magnitude of force fluctuation may vary between steps of gait (18), and this fluctuation can be exacerbated by errors due to proprioception. Clinicians should consider using low-intensity and very low-intensity resistance exercise, which are known to improve force steadiness in old adults (18, 19). These low-intensity exercise programs are appropriate precursors to moderate-intensity exercise using accentuated exercise with eccentric contractions, longer in duration or higher in load, to combat eccentric weakness in knee OA (48, 49).

One limitation of the study is that we included relatively young individuals with moderate knee OA who were also free from other medical conditions, making our subjects somewhat unusual compared with subjects in many other studies. Another limitation is that the force accuracy and steadiness measurements were assessed at relatively low force levels, and it is unknown how these force levels compare with forces that occur during ADLs. We also tested force accuracy and steadiness in a single-joint test and not in a multijoint movement, as in ADLs.

In conclusion, the present study produced 3 new findings related to individuals with knee OA. We found that knee joint proprioception error is greater in the more extended than flexed knee joint positions. We also found that OA patients have an impaired ability to accurately and steadily produce submaximal quadriceps forces. Finally, we observed a preferentially large reduction in maximal eccentric strength. In total, the data suggest that quadriceps weakness in knee OA includes not only an impaired ability to produce maximal strength, but also an impaired ability to control submaximal force. These results have implications for designing exercise and rehabilitation programs for patients with knee osteoarthritis.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

We thank graduate students Stacey Beam, Jovita Jolla, Chris Mizelle, Jill Moody, and Kim Smith for their assistance with data collection and analysis.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES