Original Article

*Correspondence to M. Van Der Esch, Jan van Breemen Institute, Center for Rehabilitation and Rheumatology, Jan van Breemenstraat 2, 1056 AB Amsterdam, The Netherlands

Received: 21 September 2005; Accepted: 6 March 2006

10.1002/art.22344  About DOI

INTRODUCTION

Osteoarthritis (OA) of the knee is a common musculoskeletal disorder ([1]). Patients with OA of the knee frequently report limitations in their ability to perform activities of daily living (functional ability), such as stair climbing, walking, and household chores ([2-4]).

Muscle strength has been shown to be a determinant of the ability to perform daily activities in patients with OA of the knee ([5][6]). Available evidence from studies on the effectiveness of muscle strengthening for knee OA demonstrates consistent improvement in ability after the intervention ([7-9]). However, the magnitude of the effect varies considerably between patients. These differences may be attributable to factors that interfere with the relationship between muscle strength and functional ability, i.e., muscle strengthening may be more effective in some patients than in others ([10][11]). Joint laxity is one factor that may contribute to this difference in efficacy.

Joint laxity is defined as the displacement or rotation of the tibia with respect to the femur in the varus-valgus direction ([10]). Joint laxity may affect the relationship between muscle strength and functional ability. However, 2 opposing hypotheses exist concerning how the relationship between muscle strength and functional ability is influenced. One hypothesis is that in patients with a high knee joint laxity, there is a stronger relationship between muscle strength and functional ability. This hypothesis is based on the assumption that in patients with high laxity, muscle activity around the knee compensates for the absence of ligamentous control due to impairments of the passive restraint system. Taking on this dual role increases the importance of muscle strength for adequate functioning, which is reflected in a stronger relationship between muscle strength and functioning. Studies in patients with anterior cruciate ligament (ACL) deficiency have shown that the loss of stability provided by ligaments and capsule can be compensated by increased muscle activity ([11][12]). The pattern of increased muscle activation was also found in patients with OA of the knee ([13]). Compared with age-matched healthy adults and young adults, patients with OA of the knee had higher muscle activity during the execution of daily activities. Therefore, in lax knee joints the role of muscle strength becomes more important, resulting in a stronger relationship between muscle strength and functional ability. The other hypothesis is that in patients with high knee joint laxity, there is a weaker relationship between muscle strength and functional ability ([14]). This hypothesis is based on the assumption that in patients with high laxity, muscle activity can no longer stabilize the knee, resulting in inadequate control of joint motion. In these patients, functional ability will be affected regardless of the level of muscle strength, resulting in a weaker relationship between muscle strength and function. In view of these 2 opposing hypotheses, the objective of this study was to establish the influence of knee joint laxity on the strength of the relationship between muscle strength and functional ability.

PATIENTS AND METHODS

Patients.

A total of 86 patients diagnosed with OA of the knee were included in the study. Inclusion criteria were OA of the knee (unilateral or bilateral), age 40-85 years, and consent to participate. Knee OA was diagnosed according to the clinical criteria of the American College of Rheumatology ([15]). Exclusion criteria were as follows: polyarthritis, presence of rheumatoid arthritis or other systemic inflammatory arthropathies, knee surgery within the last 12 months or a history of knee arthroplastic surgery, intraarticular corticosteroid injections into either knee within the previous 3 months, and/or inability to understand the Dutch language.

Measures.

Demographics.

A series of demographic variables were obtained including age, sex, height, weight, and duration of symptoms.

Muscle strength.

Muscle strength was assessed for flexion and extension of the knee using an isokinetic dynamometer (EnKnee; Enraf-Nonius, Delft, The Netherlands). Quadriceps and hamstring strength were measured isokinetically at 60°/second.

A single tester assessed all patients according to a standardized protocol. Patients were seated on a bench and secured to the testing device through the use of chest, pelvis, and thigh straps. The ankle pad of the dynamometer was placed 2 cm proximal to the medial malleolus to allow ankle dorsiflexion during the tests. The mechanical axis of the dynamometer was aligned with the approximate axis of the knee through the lateral epicondyle of the femur. Patients rested their hands on the sides of the bench.

During isokinetic testing at 60°/second, range of motion was limited to 20-80° for joint protection. Following instruction, patients performed 4 warm-up repetitions, beginning with submaximal contractions and building to maximal contractions. Following a 30-second rest, patients performed 3 maximal test repetitions. Right-left order of testing was alternated between patients. The tester verbally encouraged the patients to achieve maximal torque. The maximum score of the 3 repetitions was used to indicate maximum flexion or extension strength. The mean of flexion and extension strength of the left and right leg were computed to obtain mean muscle strength. Subsequently, mean muscle strength (in Nm) was divided by the patient's weight to control for the correlation between muscle strength and weight. Thus, a measure of overall leg muscle strength (in Nm/kg) was obtained, which was used in the analyses.

Joint laxity.

Varus-valgus laxity was measured using a previously described device and protocol that provide thigh and lower-leg immobilization, a stable knee angle in flexion of 20°, and fixed varus and valgus load ([16]). Laxity was measured (in degrees) as the movement in the frontal plane after varus and valgus load. A weight of 1.12 kg was used to load the lower leg. This weight was attached to the free-moving arm by a cord. The cord was attached 0.68 meters from the pivot of the arm, resulting in a net moment on the knee of 7.7 Nm. This load could be applied to the lower leg both medially and laterally, resulting in varus or valgus movement in the knee joint.

All measurements of laxity were performed by the same examiner (MvdE) in adherence to a protocol, including the use of anatomic landmarks for patient positioning, patient instructions, and the examiner's position. Right-left order of testing was alternated between patients. Three consecutive measurements were made. The mean (in degrees) laxity of the right and left knees obtained from these 3 measurements was used for analysis. The intraclass correlation coefficients (ICCs) for intrarater and interrater reliability of the measurements with this device in healthy persons were 0.80 and 0.88, respectively ([17]).

Functional ability.

Functional ability was assessed with both a standardized physical performance test and a self-report questionnaire (Western Ontario and McMaster Universities Osteoarthritis Index [WOMAC]). As a performance-based measure of function, a 100-meter walking test was used ([18]). The time to walk a 20-meter level and unobstructed corridor 5 times (100 meters in total) was measured. Patients were instructed to walk the distance as fast as possible. On the command  go,  patients walked along the level of the corridor. They were instructed not to stop before crossing the finish line. A stopwatch was used to measure in seconds the time from the command  go  until patients crossed the finish line. The examiner was standing at the finish line during the test. Patients who used canes while walking were permitted to use them during the test. All patients were wearing walking shoes.

The Dutch version of the WOMAC was used to assess self-reported functional ability ([19]). The WOMAC is a disease-specific measure of pain, stiffness, and physical function for individuals with OA of the knee ([20]). The WOMAC, with a possible score range of 0-96, includes 5 items related to pain, 2 items related to stiffness, and 17 items related to physical function. Each item is scored on a 5-point Likert scale. Reliability and validity of the WOMAC have been established ([21]). Higher scores on the WOMAC represent greater limitations in function. The ICC for Dutch WOMAC physical functioning was 0.92 ([19]).

Pain.

Average overall pain in the past week and average current knee pain were measured using a 100-mm visual analog scale.

Radiography.

Radiographs of the knee were scored by an experienced reader using the grading scales proposed by Kellgren and Lawrence ([22][23]). The radiographs of 7 patients were missing.

Statistical analysis.

Because functional ability (i.e., walking ability and WOMAC physical function score) was specific to the person, knee-specific data (i.e., strength and joint laxity) were averaged across right and left knees for analyses involving functional ability.

First, Pearson's correlation coefficients were computed to establish the bivariate relationship between joint laxity and muscle strength and between joint laxity and functional ability (i.e., walking time and WOMAC physical function). Second, multiple regression analyses were performed to assess the relationship between muscle strength and functional ability and the impact of laxity. Multiple regression analyses were used to assess which factors were independently associated with functional ability. An interaction variable between muscle strength and laxity was added to the model to assess the role of laxity as a modifier of the relationship between muscle strength and functional ability. The independent variables muscle strength and joint laxity were centered around the mean ([24]). Centering allows for a meaningful interpretation of main effects when interaction is present in the model. Other independent variables in the analysis comprised age, sex, duration of symptoms, and current pain. The significance level for exclusion from the final regression model was set at P < 0.10; regression coefficients were considered to be significant at P < 0.05. All analyses were performed using SPSS software, version 11.5 (SPSS, Chicago, IL).

RESULTS

Characteristics of the study sample are listed in Table 1. The mean varus-valgus laxity between the left and right knees correlated with each other (r = 0.78, P < 0.001). Between the left and right knees, quadriceps strength and hamstring strength correlated with each other (r = 0.79, P < 0.001 and r = 0.83, P < 0.001, respectively). The mean ± SD total muscle strength as an average of flexion and extension strength was 0.74 ± 0.35 Nm/kg, with a Pearson's correlation coefficient of 0.85 (P < 0.001) between the average of quadriceps and hamstring strength of the left knee and the average of quadriceps and hamstring strength of the right knee.

Bivariate relationships between joint laxity, muscle strength, and functional ability.

Joint laxity was moderately associated with walking time (r = 0.25, P < 0.05) and not associated with the WOMAC physical function score (r = 0.03, P = 0.799). Negative correlations were found between joint laxity and total muscle strength (r = -0.34, P < 0.05) and between total muscle strength and walking time (r = -0.50, P < 0.001). Similarly, total muscle strength correlated negatively with the WOMAC physical function score (r = -0.61, P < 0.001).

Multivariate relationships between joint laxity, muscle strength, and functional ability.

To analyze the relationship between functional ability and total muscle strength, a multiple regression model was constructed:

The model explaining the total variation of walking time was as follows (see Table 2): walking time = 97.41 - 72.73 muscle strength + 0.70 laxity - 12.24 muscle strength × laxity (F = 13.89, P < 0.001, R² = 0.35; N = 81). This means that 35% of the total variation of walking time is explained by muscle strength, laxity, and their interaction. The independent variable muscle strength (b₁ = -72.73, P < 0.001) and the interaction between muscle strength and joint laxity (b₃ = -14.24, P = 0.002) were significantly associated with walking time. When laxity equals 0 (0 = mean of 6.9°) and muscle strength increases by 1 Nm/kg, then the walking time will decrease by 72.73 seconds. However, when laxity increases by 1° (1 = 7.9°) and muscle strength increases by 1 Nm/kg (1 = 1.74 Nm/kg), then the walking time will decrease by 84.27 seconds.

The model explaining the total variation of WOMAC physical function was as follows (see Table 2): WOMAC physical function = 30.98 - 31.49 muscle strength - 1.04 laxity - 2.34 laxity × muscle strength (F = 19.94, P < 0.001, R² = 0.43; N = 81). This means that 43% of the total variation of WOMAC physical function is explained by muscle strength, laxity, and their interaction. The independent variables muscle strength (b = -31.49, P < 0.001), joint laxity (b = -1.04, P < 0.05), and the interaction between these 2 variables (b = -2.34, P = 0.08) were associated with the WOMAC physical function score, although the interaction was not statistically significant at the P < 0.05 level. This means that when laxity equals 0 (0 = mean of 6.9°) and muscle strength increases by 1 Nm/kg (0 = mean of 0.74 Nm/kg), then the WOMAC physical function score will decrease by 21.48. However, when laxity increases by 1° (1 = 7.9°) and muscle strength increases by 1 Nm/kg, then the WOMAC physical function will decrease by 34.87. To visualize the interaction between muscle strength and joint laxity, laxity was dichotomized into low and high laxity using the median-split method (Figure 1).

These analyses were repeated in a more extensive model, with the demographic variables from Table 1 as controlling variables. The results of those analyses were consistent with the results published here.

DISCUSSION

Two opposing hypotheses of the influence of joint laxity on the relationship between muscle strength and functional ability were tested in patients with OA of the knee. Our results confirm the first hypothesis, i.e., high joint laxity is associated with a stronger relationship between muscle strength and functional ability.

The results of the present study may be explained by the results presented by Hortobagyi et al ([13]). In that study, patients with OA had a significantly higher coactivity of knee muscles than age-matched healthy adults and young adults. Patients with knee OA revealed increased muscle coactivation while executing activities of daily living. Coactivation is considered to provide active stabilization of the knee in the absence of adequate stabilization by the passive restraint system (ligaments and capsule) ([11][12]). It is likely that coactivation of muscles will only succeed in stabilizing the knee joint when there is sufficient muscle strength. This means that muscle strength is a prerequisite for successful joint stabilization through muscle coactivation. Therefore, we hypothesize that coactivation will be more successful in providing joint stability and subsequently in maintaining functional ability in patients with high muscle strength than in patients with less muscle strength, indicating a close relationship between available muscle strength and successful stabilization through muscle coactivation. This would mean that in patients with high knee joint laxity, differences in muscle strength result in relatively large differences in functional ability compared with patients with low-laxity knee joints. Comparable results were found in a study by Doorenbosch and Harlaar ([11]), where subjects with an ACL deficiency, i.e., high anterior-posterior laxity, compensated the loss of passive stability (laxity) by developing higher coactivation levels of knee muscles, i.e., active stabilization. Similarly, the results of McNair and Marshall ([12]) support the hypothesis that higher levels of co-contraction of quadriceps and hamstrings during movements in patients with ACL deficiency provide an active stabilization of the knee to compensate for the loss of the passive structure.

Our results are not in agreement with conclusions presented by Sharma et al ([9][10][14]). In one of those studies ([14]), it was stated that high laxity was associated with a weaker relationship between muscle strength and functional ability in patients with knee OA (supporting the second hypothesis). A likely explanation of this discrepancy is the difference in analytical approach. The conclusions of Sharma et al were based on a comparison of the correlations between muscle strength and disability in a high-laxity group and low-laxity group. Between these 2 patient groups, there was a small difference in correlation between quadriceps strength and WOMAC physical function (r = -0.27, 95% confidence interval [95% CI] -0.46, -0.05 in the low-laxity group and r = -0.19, 95% CI -0.40, 0.04 in the high-laxity group) and between hamstring strength and WOMAC physical function (r = -0.30, 95% CI -0.39, 0.03 in the low-laxity group and r = -0.21, 95% CI -0.42, 0.02 in the high-laxity group). Given these 95% CIs, it is not likely that the differences in correlation reported by Sharma et al were statistically significant. Additionally, for our particular research question, the use of regression coefficients is preferable. First, using a regression model with an interaction term of muscle strength and laxity allows for one analysis using data from all patients, whereas a correlational analysis similar to the approach used by Sharma et al would require dividing the research group into patients with high and low laxity based on an arbitrary cutoff point. Second, the P value of the regression coefficient of the interaction term provides an immediate insight into the statistical significance of the impact of laxity on the relationship between muscle strength and functional ability.

It should be noted that there are some differences between the populations and measurement equipment and protocols of our study and the study by Sharma et al ([14]). Our patients were on average more disabled (higher WOMAC physical function score), although age, sex, body mass index, pain, and OA severity were similar. With regard to the measurement protocols and equipment, there were differences between the studies in measuring laxity and muscle strength. In our measurement of laxity, we applied a different method of leg fixation to the device; used a lower torque, which was also applied in a different manner; and used an electronic sensor to assess varus-valgus rotation rather than an analogous device. In our study, muscle strength was measured isokinetically with a lower velocity (60°/second as opposed to 120°/second). Muscle strength was also corrected for body weight and was expressed in SI units rather than feet/pound. However, although these differences may have influenced the results, we believe that the statistical analysis is the main reason for the different conclusions.

The direct relationship between laxity and functional ability was found to be weak (walking time) or absent (WOMAC physical function). Therefore, although laxity is an important factor in instability of the knee ([25]), the direct effect of laxity in functional ability seems to be relatively limited.

Some issues need to be addressed concerning the methods used in this study. First, the interrelationship of joint laxity between left and right knees in patients with OA of the knee was established and showed a high correlation. Consequently, joint laxity of the left and right knees of the same patients was averaged and used in subsequent analyses. Second, the interrelationship of muscle strength between the left and right knees was established, also showing a high correlation. The results of the muscle strength measurements were averaged in the same manner and were used in subsequent analyses. This indicated that both knee joint laxity and muscle strength are characteristics of a specific patient, instead of characteristics of a specific knee. This finding has been reported previously for muscle strength ([5]).

In considering the implications of this study for exercise therapy, it is useful to consider some limitations first. One limitation is that an adequate level of joint laxity is unknown. In the absence of a known cutoff point to separate normal angular deviation under load from abnormal deviation (laxity), the differentiation between high and low laxity is only relative. The second limitation of this study is that it was a cross-sectional study of 86 patients from 1 rehabilitation center and causal conclusions were not allowed. Nevertheless, our results support the use of exercise therapy in patients with OA with high knee joint laxity. Based on the results presented here, patients with high laxity can be predicted to benefit from interventions aimed at increasing muscle strength.

In conclusion, patients with OA with high knee joint laxity show a stronger relationship between muscle strength and functional ability than patients with OA with low knee joint laxity. Patients with OA with high knee joint laxity and low muscle strength are most at risk of being disabled.

Acknowledgements

We gratefully acknowledge Ms K. Fiedler for her assistance in correcting the manuscript.