SEARCH

SEARCH BY CITATION

Abstract

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

Objective

Knee extensor strength is an important correlate of physical function in patients with knee osteoarthritis; however, it remains unclear whether standing balance is also a correlate. The purpose of this study was to evaluate the cross-sectional associations of knee extensor strength, standing balance, and their interaction with physical function.

Methods

One hundred four older adults with end-stage knee osteoarthritis awaiting a total knee replacement (mean ± SD age 67 ± 8 years) participated. Isometric knee extensor strength was measured using an isokinetic dynamometer. Standing balance performance was measured by the center of pressure displacement during quiet standing on a balance board. Physical function was measured by the self-report Short Form 36 (SF-36) questionnaire and by the 10-meter fast-pace gait speed test.

Results

After adjustment for demographic and knee pain variables, we detected significant knee strength by standing balance interaction terms for both SF-36 physical function and fast-pace gait speed. Interrogation of the interaction revealed that standing balance in the anteroposterior plane was positively related to physical function among patients with lower knee extensor strength. Conversely, among patients with higher knee extensor strength, the standing balance–physical function associations were, or tended to be, negative.

Conclusion

These findings suggest that although standing balance was related to physical function in patients with knee osteoarthritis, this relationship was complex and dependent on knee extensor strength level. These results are of importance in developing intervention strategies and refining theoretical models, but they call for further study.


INTRODUCTION

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

Limitations in physical function are associated with participation restrictions in people with knee osteoarthritis (OA) (1), and they predicted mortality independent of traditional risk factors (2). Compared with people without knee OA, those with knee OA have lower knee extensor strength (force-generating capacity) (3–6) and impaired standing balance (5–8), commonly defined as greater postural body sway or center of pressure (COP) displacement, suggesting that both reduced knee extensor strength and standing balance are potential impairments to be evaluated in the research and clinical settings as they may influence functional limitations in knee OA. Although several observational studies have shown with reasonable consistency that knee extensor strength is an important determinant of physical function (4, 9), evidence for the standing balance–physical function association is limited (4, 5, 7). In the first place, it is not a universal observation that patients with knee OA had greater standing COP displacements than did healthy controls (3, 4, 10). Furthermore, 3 cross-sectional studies (4, 5, 7) in patients with knee OA reported no significant association between standing balance and physical function in either bivariate or multivariable analyses. Of these studies, Hurley and colleagues (4) performed a subgroup analysis of 23 patients who were unable to maintain a bipedal stance for more than 7 seconds, and the authors found a weak bivariate association (r = 0.31) between standing balance and physical function.

We believe the null findings (weak or no balance effects) of previous studies may reflect the underlying complexity of adequately examining the association between standing balance and physical function. Indeed, while knee strength and standing balance have been separately examined in relation to physical function, no studies have examined their possible interactive effects on function. And yet, this possibility merits exploration for at least 2 reasons. First, knee extensor strength and standing balance impairments were moderately associated with each other (Spearman's ρ = 0.40) (5) and tended to co-occur in patients with knee OA. Second, it should be remembered that standing balance performance is a multivariate phenomenon (11) and therefore, although the use of COP displacement as a measure of standing balance is firmly established in the literature, the interpretation of the COP trajectory requires thoughtful consideration (12). Specifically, the COP position reflects the “net neuromuscular response to control of the passive center of gravity” (11), and it is biologically plausible for different neuromuscular strategies to produce similar levels of COP displacement (12, 13) with possibly divergent levels of physical function. To the extent that both knee extensor strength and standing balance have been linked to these neurophysiologic processes (see Discussion), a failure to examine their combined interactive effects may contribute to the inconsistent nature of previous findings.

Therefore, the question remains whether and how knee extensor strength moderates the relationship between standing balance and physical function in knee OA. As we believed that this information could potentially assist the refinement of theoretical models and the development of effective interventions, we initiated the present study to explore the cross-sectional associations of knee extensor strength, standing balance, and their interaction with physical function in a sample of older adults with knee OA.

Significance & Innovations

  • Standing balance, indexed by the center of pressure (COP) displacement in the anteroposterior plane, interacted with knee extensor strength to influence physical function in patients with knee osteoarthritis (OA).

  • The interactions were disordinal in form: standing balance was positively related to physical function among patients with lower knee extensor strength but was, or tended to be, negatively related to physical function among patients with higher strength levels.

  • It is possible that the inconsistent findings of past studies may have, at least in part, resulted from not considering the level of knee strength.

  • The present results also suggest that the interpretation of standing COP measures in patients with knee OA is certainly less straightforward than previously assumed.

PATIENTS AND METHODS

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

Patients.

Our study sample comprised 104 patients with end-stage knee OA undergoing unilateral total knee replacement at a large tertiary institution in Singapore from June 2010 to January 2011. Patients were recruited within a month before their surgery as part of a randomized clinical trial investigating the effects of postoperative electrical muscle stimulation. The present study is concerned with the baseline data obtained prior to surgery. Patients were excluded if they 1) were unable to walk 10 meters independently without an assistive device, 2) had significant back or other non-knee pain, 3) had secondary knee OA due to trauma or inflammatory or metabolic rheumatic diseases, or 4) had any medical conditions that would compromise physical function or affect their ability to complete testing. The study was approved by the SingHealth Institutional Review Board and was conducted in accordance with the Declaration of Helsinki.

Study procedure.

Patients attended a test session at the outpatient physiotherapy department after providing written informed consent. Prior to the physical assessment, patients' height and weight were obtained. Each patient also completed the Short Form 36 (SF-36) general health survey (14) for physical function and bodily pain subscales. The SF-36 is a generic questionnaire designed to measure health-related quality of life in general and specific populations (15). It comprises 36 items related to 8 subscales of health, of which we used the bodily pain and physical function subscales. Each subscale ranges from 0–100, with higher scores representing better health states. The English and Chinese versions of the SF-36 have been validated for use in Singapore (16).

Physical measures.

Testing was performed in the following order: standing balance test, gait speed test, and knee extensor strength test of the knee awaiting replacement surgery. Patients were provided with rest periods as requested. Of note, the order of the physical measures was not randomized and the least strenuous test was placed before the most strenuous test in order to reduce the potential confounding influence of pain.

Standing balance test.

Standing balance was assessed using the Wii Balance Board (Nintendo) in a protocol previously validated against a laboratory force plate (17). The patients stood unsupported on the Wii Balance Board in their usual comfortable stance, and were instructed to keep their hands by their side, look straight ahead, and stand quietly for the duration of the trial. Two trials of a duration of 30 seconds each were performed.

The Wii Balance Board was interfaced with a laptop computer using custom-written software (Labview 8.5 National Instruments), and was calibrated by placing a variety of known loads at different positions on the Wii Balance Board (17). Anteroposterior (AP) and mediolateral (ML) COP coordinates were recorded at 40 Hz and low-pass filtered at 12 Hz using an eighth-order Butterworth filter to eliminate noise. The range and SD of the COP trajectory along both AP and ML planes were calculated because these measures are common in the extant knee OA literature. The range of the COP represented the distance between the most positive and negative COP trajectory positions in the respective planes; the SD of the COP represented the variability of the COP around the mean position in the respective planes. For each standing balance measure, the mean of 2 trials was used.

Gait speed test.

For the gait speed test, the patients were timed using a stopwatch as they walked along a 10-meter walkway at a fast pace. Patients stood directly behind the start line and were clocked from the time the first foot crossed the start line until the lead foot crossed the finish line. Patients were instructed to “walk as quickly as you can, but safely” and to finish at least 2 meters past the finish line to eliminate the deceleration effects from stopping the walk. Each patient performed 2 valid trials and the better trial was taken for analysis (18).

Knee extensor strength.

Maximal volitional isometric contractions of the knee extensors at 75° of knee flexion were performed on a Biodex Medical IV isokinetic dynamometer. The patient was tested in a seated position with the hip at 90° of flexion, and strapping was placed across the waist and chest to stabilize the torso. The axis of rotation of the dynamometer lever arm was aligned to the femoral lateral condyle, and the lever arm was secured to the tibia just proximal to the medial malleolus via an ankle cuff. Before testing the knee extensors, the gravity compensation procedure was performed by measuring the patient's passive extremity weight at 30° of knee flexion. Following a warm-up comprising 1 submaximal and 1 maximal contraction, all of the patients performed 2 maximal trials for 5 seconds with a 1-minute rest interval. Patients were instructed to extend their knee as fast and forcefully as possible for at least 2 to 3 seconds, and additional measurements were taken if the patients reported a failure to achieve a maximum effort. Standardized strong verbal encouragement was given during testing and visual feedback was provided after each trial. In each trial, the peak torque normalized to the patient's body mass was recorded and the higher measurement of 2 valid trials was analyzed.

Statistical analyses.

We used descriptive statistics to characterize the study sample: we used means with SDs or medians with interquartile ranges (IQRs) for continuous variables and frequencies with percentages for categorical variables. Normality was assessed visually and tested using the Kolmogorov-Smirnov method. To examine the interrelations between the explanatory variables, we performed agglomerative hierarchical variable clustering with the squared Spearman's rank correlation as the similarity matrix (19).

Separate hierarchical linear regression models were used to examine the associations of standing balance, knee extensor strength, and their interaction with the gait and SF-36 physical function measures. In step 1, we entered into each hierarchical model the standing balance and knee extensor strength measures and the covariates that were chosen based on their potential or documented (9) association with physical function, specifically, sex, age, body weight, and SF-36 bodily pain. In step 2, we entered the cross-product interaction term for standing balance and knee extensor strength. When significant interaction effects were found, we applied the Johnson-Neyman technique using the MODPROBE macro (20) to determine the range of knee extensor strength values over which the conditional slopes between standing balance and physical function were statistically significant. Finally, to be cautious about overinterpreting the results from the Johnson-Neyman analyses, we operationally defined patients as having poor gait speed if the fast-pace gait speed fell below 1.0 meters/second (21), and the receiver operating characteristic (ROC) curve and the Youden index (22) were used to identify a knee strength cut point for the definition of poor gait speed.

Statistical analyses were done with R software, version 2.12.1 (R Foundation), using the Hmisc/rms and pROC packages (available online at http://lib.stat.cmu.edu/R/CRAN/). Statistical significance was determined at the 2-sided 0.05 level.

RESULTS

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

Table 1 shows the patients' sociodemographic and functional characteristics. Women accounted for nearly three-quarters (73%) of the sample, and the mean ± SD body mass index for all of the patients was 27.3 ± 4.9 kg/m2. Forty-two percent of the patients were classified as having a poor gait speed (<1.0 meters/second). Figure 1 shows the correlational structure of our explanatory variables. The standing COP variables aggregated into the ML and AP planes and the COP range and SD variables within each plane were highly colinear (Spearman's ρ2 = ≥0.90), thereby justifying the use of separate regression models.

Table 1. Characteristics of patients (n = 104)*
 Value
  • *

    Values are the mean ± SD unless otherwise indicated. BMI = body mass index; IQR = interquartile range; SF-36 = Medical Outcomes Study 36-item Short Form health survey; ML = mediolateral; AP = anteroposterior.

Age, years67.4 ± 7.6
Female sex, no. (%)76 (73)
Height, meters1.56 ± 0.08
Body mass, kg66.2 ± 13.6
BMI, kg/m227.3 ± 4.9
Duration of symptoms, median (IQR) years5 (8)
Bilateral symptoms, no. (%)68 (65)
Knee extensor strength, Nm/kg1.13 ± 0.49
SF-36 bodily pain42.2 ± 21.5
Standing balance measures 
 Range ML, cm1.62 ± 0.65
 SD ML, cm0.33 ± 0.14
 Range AP, cm2.42 ± 0.68
 SD AP, cm0.48 ± 0.14
Physical function measures 
 Gait speed, meters/second1.03 ± 0.31
 Poor gait speed (1.0 meter/second),  no. (%)44 (42)
 SF-36 physical function41.1 ± 20.6
thumbnail image

Figure 1. Dendrogram of the hierarchical clustering of the explanatory variables, using squared Spearman's rank correlation as the similarity matrix. The horizontal lines at which variables connect represent the squared correlation coefficient, and a larger Spearman's ρ2 value indicates greater correlation. ML = mediolateral; AP = anteroposterior; SF-36 = Medical Outcomes Study 36-item Short Form health survey.

Download figure to PowerPoint

Tables 2 and 3 show the hierarchical regression models for fast-pace gait speed and SF-36 physical function, respectively. The combination of knee extensor strength, standing COP variables, and covariates explained between 32% and 55% of the variation in physical function. Across the models, age and SF-36 bodily pain were consistently associated with physical function. Knee extensor strength was a significant correlate for gait speed and not the SF-36 physical function model, whereas the main effect (i.e., an effect not conditional on other variables) for the standing COP measures was observed only for the COP SD AP measure in the gait speed model (b = 0.38, P = 0.02).

Table 2. Multivariable models of the association of fast-pace gait speed (the dependent variable) with standing balance measures (n = 104)*
 β ± SEP
  • *

    Beta values are unstandardized regression coefficients. SF-36 = Medical Outcomes Study 36-item Short Form health survey; ML= mediolateral; AP = anteroposterior.

  • P < 0.001.

  • 1 = men, 2 = women.

Step 1 (adjusted R2 = 0.52)  
 Age−0.021 ± 0.003< 0.0001
 Sex−0.201 ± 0.059< 0.001
 Body mass−0.0038 ± 0.0020.04
 SF-36 bodily pain0.0025 ± 0.0010.02
 Knee extensor strength0.13 ± 0.0540.02
 Range ML−0.056 ± 0.030.09
Step 2  
 Knee strength × range ML0.028 ± 0.0710.69
Step 1 (adjusted R2 = 0.52)  
 Age−0.021 ± 0.003< 0.0001
 Sex−0.198 ± 0.059< 0.01
 Body mass−0.0037 ± 0.0020.04
 SF-36 bodily pain0.0026 ± 0.0010.02
 Knee extensor strength0.13 ± 0.0530.02
 SD ML−0.28 ± 0.160.08
Step 2  
 Knee strength × SD ML0.246 ± 0.3460.48
Step 1 (adjusted R2 = 0.53)  
 Age−0.023 ± 0.003< 0.0001
 Sex−0.18 ± 0.06< 0.01
 Body mass−0.0037 ± 0.0020.04
 SF-36 bodily pain0.0027 ± 0.0010.01
 Knee extensor strength0.138 ± 0.0540.01
 Range AP0.059 ± 0.0320.07
Step 2  
 Knee strength × range AP−0.14 ± 0.080.09
Step 1 (adjusted R2 = 0.55)  
 Age−0.023 ± 0.003< 0.0001
 Sex−0.17 ± 0.06< 0.01
 Body mass−0.0036 ± 0.0020.04
 SF-36 bodily pain0.0028 ± 0.001< 0.01
 Knee extensor strength0.14 ± 0.050.01
 SD AP0.38 ± 0.160.02
Step 2  
 Knee strength × SD AP−0.92 ± 0.420.03
Table 3. Multivariable models of the association of SF-36 physical function (the dependent variable) with standing balance measures (n = 104)*
 β ± SEP
  • *

    Beta values are unstandardized regression coefficients. SF-36 = Medical Outcomes Study 36-item Short Form health survey; ML= mediolateral; AP = anteroposterior.

  • P < 0.001.

  • 1 = men, 2 = women.

Step 1 (adjusted R2 = 0.33)  
 Age−0.64 ± 0.250.01
 Sex−9.06 ± 4.60.05
 Body mass−0.12 ± 0.140.38
 SF-36 bodily pain0.52 ± 0.082< 0.0001
 Knee extensor strength1.79 ± 4.20.67
 Range ML0.38 ± 2.610.89
Step 2  
 Knee strength × range ML−9.28 ± 5.60.10
Step 1 (adjusted R2 = 0.32)  
 Age−0.65 ± 0.30.01
 Sex−9.13 ± 4.660.05
 Body mass−0.12 ± 0.140.37
 SF-36 bodily pain0.52 ± 0.08< 0.0001
 Knee extensor strength1.79 ± 4.230.67
 SD ML4.08 ± 12.30.74
Step 2  
 Knee strength × SD ML−37.5 ± 27.20.17
Step 1 (adjusted R2 = 0.42)  
 Age−0.7 ± 0.2< 0.01
 Sex−7.49 ± 4.670.11
 Body mass−0.12 ± 0.140.38
 SF-36 bodily pain0.52 ± 0.08< 0.0001
 Knee extensor strength2.83 ± 4.210.50
 Range AP4.33 ± 2.510.09
Step 2  
 Knee strength × range AP−23.3 ± 6< 0.001
Step 1 (adjusted R2 = 0.42)  
 Age−0.68 ± 0.25< 0.01
 Sex−7.23 ± 4.690.13
 Body mass−0.12 ± 0.140.40
 SF-36 bodily pain−0.52 ± 0.08< 0.0001
 Knee extensor strength2.51 ± 4.180.55
 SD AP22.3 ± 12.40.08
Step 2  
 Knee strength × SD AP−121.2 ± 31.4< 0.001

The moderational effect of knee extensor strength was evident for the association between physical function and standing COP AP variables, but not COP ML variables. Figure 2 depicts these interactive effects by showing the association between COP SD AP and physical function at different levels of knee extensor strength. Specifically, the Johnson-Neyman analyses (Table 4) indicate that at levels of knee extensor strength below 1.10 Nm/kg and 1.01 Nm/kg, greater standing COP SD AP values were significantly related to greater (better) fast-pace gait speed and SF-36 physical function, respectively. However, the strength of the gradients diminished as knee strength levels increased, and at levels greater than 1.50 Nm/kg, greater standing COP SD AP values were significantly inversely related to lower (poorer) SF-36 physical function. As a secondary analysis, we performed an ROC analysis to determine the knee extensor cut point that discriminated patients with and without poor gait speed, and the optimal cut point based on the Youden index was 1.22 Nm/kg (area under the ROC curve 0.74, 95% confidence interval 0.64, 0.83).

thumbnail image

Figure 2. Three-dimensional plots showing the influence of standing center of pressure (COP) SD anteroposterior (AP), knee extensor strength, and their interaction on fast-pace gait speed (A) and Medical Outcomes Study 36-item Short Form health survey (SF-36) physical function (B), adjusted for sex, age, body weight, and SF-36 bodily pain.

Download figure to PowerPoint

Table 4. Johnson-Neyman significance regions for standing center of pressure SD anteroposterior predicting physical function at a range of values of knee extensor strength*
Knee extensor strength (Nm/kg)bSEP95% CI
  • *

    95% CI = 95% confidence interval; SF-36 = Medical Outcomes Study 36-item Short Form health survey.

  • Knee extensor strength value where the effect of standing balance on physical function transitions between statistically significant and nonsignificance.

Fast-pace gait speed    
 0.15 (−2.0 SDs)1.180.400.0040.40, 1.96
 0.40 (−1.5 SDs)0.960.300.0020.35, 1.56
 0.64 (−1.0 SD)0.730.220.0010.29, 1.17
 0.89 (−0.5 SD)0.540.170.0020.20, 0.88
 1.10 (−0.01 SD)0.310.160.050.0, 0.62
 1.13 (mean)0.280.160.08−0.04, 0.60
 1.38 (+0.5 SD)0.050.210.80−0.37, 0.48
 1.62 (+1.0 SD)−0.170.290.57−0.75, 0.41
 1.87 (+1.5 SDs)−0.420.390.29−1.20, 0.36
 2.11 (+2.0 SDs)−0.630.490.19−1.59, 0.33
SF-36 physical function    
 0.15 (−2.0 SDs)127.729.8< 0.00168.5, 186.9
 0.40 (−1.5 SDs)98.223.0< 0.00152.6, 143.7
 0.64 (−1.0 SD)68.516.8< 0.00135.2, 101.8
 0.89 (−0.5 SD)43.312.90.00117.7, 68.9
 1.01 (−0.23 SD)23.111.640.050.0, 46.2
 1.13 (mean)9.212.10.45−14.8, 33.3
 1.38 (+0.5 SD)−21.016.20.20−53.1, 11.1
 1.53 (+0.8 SD)−39.219.740.05−78.4, 0.0
 1.62 (+1.0 SD)−50.122.10.03−93.9, −6.2
 1.87 (+1.5 SDs)−83.329.80.006−142.4, −24.2
 2.11 (+2.0 SDs)−111.436.60.003−184.1, −38.7

DISCUSSION

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

To our knowledge, never before has the interaction between standing balance and knee extensor strength on physical function been examined in patients with knee OA. In doing so, we sought to reconcile previously reported null and equivocal findings concerning associations between standing balance and physical function in knee OA. The interactions we found were to be disordinal in form (Figure 2): COP AP variables were positively related to physical function among patients whose knee extensor strength fell below 1.10 Nm/kg but were, or tended to be, negatively related to physical function among patients with higher strength levels. To interpret these results in a more functional context, the ROC analysis showed that the optimum knee extensor strength threshold for separating patients with and without a poor gait speed was 1.22 Nm/kg. Because this cut point exceeded the cut points derived from the Johnson-Neyman analyses, our findings collectively suggest that COP AP displacement variables were only positively related to physical function in circumstances of substantial knee extensor weakness.

Before a discussion of data interpretation, it is noteworthy that statistically significant balance by strength interaction effects were observed in relation to AP, but not ML, variables of standing balance. In our study, hierarchical variable clustering (Figure 1) revealed that the correlation between the ML and AP variables was modest (Spearman's ρ2 = <0.20); therefore, although derived from the same test, standing COP ML and AP variables were separable and thus potentially capable of influencing physical function independently. Furthermore, this finding reinforces the notion that the central nervous system (CNS) uses different strategies to control balance in the AP and ML planes (23) and given also that the knee extensors are primarily sagittal plane movers, it seems permissible to suggest that the knee extensors may be more involved in postural control in the AP plane than in the ML plane (23, 24).

Turning to the results for the COP AP variables, we found that among patients without substantial knee extensor weakness (>1.10 Nm/kg), lesser standing COP AP excursion tended to be associated with better physical function (Figure 2). These findings are consistent with an earlier study by Hurley et al (4) demonstrating a weak bivariate association between standing balance and function in 23 patients with knee OA, and with the conventional assumption that presumes that lesser COP displacement is reflective of better postural control. As mentioned in the Introduction, standing balance performance is a multivariate phenomenon of which knee muscle performance is one of the many determinants (25). Accordingly, in the absence of substantial knee extensor weakness, we reasoned, as did Hurley et al (4), that any deficiencies within the sensory, integrative, and motor components of the postural control system may increase COP AP excursion (26–28) and adversely affect physical function.

Conversely, among patients with substantial knee extensor weakness (<1.10 Nm/kg), greater COP AP excursion was associated with better physical function. Although departing from the conventional assumption that lesser COP displacement is associated with better postural control and physical function, our results are biologically explainable. Specifically, our results should be viewed in the context of the accumulating evidence (29–31) suggesting that COP excursion is an exploratory activity of the CNS to ensure an adequate level of sensory input. Given that skeletal muscles have important motor and sensory (proprioceptive) functions (32), the CNS may, in the presence of substantial muscle weakness (and the associated sensory deficits), adopt at least 2 compensatory strategies to achieve standing balance. First, the CNS may purposefully increase COP excursion in an attempt to use the rich sensory information from the remaining sensory organs; hence, this strategy would link greater AP sway with better physical function among patients with substantial muscle weakness. The other strategy enlists global muscle co-contraction to reduce COP excursion, and has sometimes been referred to as a postural stiffening strategy (33, 34).

Reviewing the literature, the most relevant evidence that patients with knee OA may adopt a co-contraction strategy during quiet standing derives from the work by Lewek et al (35), who observed greater medial muscle co-contraction in patients with knee OA compared with controls with healthy knees. Furthermore, numerous reports have observed the co-contraction of various lower extremity muscles in knee OA gait (34, 36–41), and importantly, this co-contraction strategy was manifest particularly among patients with substantial knee extensor weakness (34, 41). Therefore, taken together, it seems likely that a subgroup of our patients with substantial knee extensor weakness may adopt the postural stiffness strategy during quiet standing.

If this were true, one question to be asked of this strategy is whether it is maladaptive. Although co-contraction of the lower extremity muscles may indeed reduce postural sway during standing (42–44), our results indicate that the combination of a small COP excursion and knee extensor weakness was associated with a low level of physical function. Schmitt and Rudolph (45) reported on 20 patients with medial knee OA and found that higher knee muscle co-contraction was associated with greater self-report knee instability, an important multivariate predictor of functional limitations (46), and the authors concluded that that the co-contraction strategy may be detrimental to both knee cartilage integrity and physical function. Also, biomechanical studies (33, 47) have shown that individuals with increased ankle stiffness during standing were unable to adequately counteract rapid external perturbations and that postural ankle stiffness during standing was greater in elderly fallers than in nonfallers (48). Granted, we acknowledged that our explanations are speculative and additional studies are needed to test these possibilities; however, our results suggest that the interpretation of standing COP displacement measures in patients with knee OA is certainly less straightforward than previously assumed.

Thus far, although we are inclined to invoke purely neurophysiologic explanations for our findings, it is possible that there is a psychological basis to them. For instance, fear of falling is one psychological factor that has been demonstrated to increase muscle activation and postural stiffness in the AP direction (44, 49). Given that fear of falling was associated with both knee strength and physical function in older adults with knee OA (10), it seems reasonable to suggest that a fear of falling may induce a postural stiffening strategy in some of our patients. Unfortunately, we did not have questionnaire data on falls-related self-efficacy, which would have been a useful adjunct to evaluate whether fear of falling is a confounder or a potential pathway through which knee extensor strength moderated the standing balance–function association.

Our study has implications. First, the interactive associations observed in this study raise the intriguing possibility that the inconsistent findings of past studies may have, at least in part, resulted from not considering the level of knee strength. Indeed, in samples that include patients with a broad range of muscle strength such as the present sample, it may be difficult to either detect significant main effects of standing balance or to validly compare people with or without knee OA on standing balance performance. Second, our data provide provisional evidence of an important heterogeneity in knee OA and hence, one potential avenue for future epidemiologic research may entail the disaggregation of patients with knee OA into distinct subgroups. Specifically, if it could be shown that patients with knee extensor weakness and reduced standing COP excursion have the greatest radiographic or functional deterioration, these patients may constitute a subpopulation in which early and intensive management is warranted. Third, from the clinical standpoint, although our results generally support a strengthening intervention as the mainstay of knee OA rehabilitation, clinicians should be sensitized to patients who have concomitant knee extensor weakness and reduced COP excursion during quiet standing. Management of these patients may include cognitive and perturbation training, and the latter has been shown to improve knee kinematics and muscle synergy in patients with anterior cruciate ligament deficiency (50). Again, more research is necessary before definitive clinical recommendations can be made.

Our study has limitations. First, our results may be affected by reverse causation because of their cross-sectional nature, and intervention studies are needed to establish the direction of causation. Second, we studied patients with end-stage knee OA and it is uncertain if our findings are invariantly applicable to patients with mild to moderate knee OA. Similarly, on average, our sample had low levels of knee extensor strength and physical function and thus, our study is limited in its ability to fully examine the standing balance–physical function association across a wide continuum of knee strength. Still, the fact that we were able to detect significant interaction effects underscores the robustness of our findings. Second, although our results shed light on possible mechanistic relationships in a way that looking at main effects does not, they offer little insight into exactly why knee extensor strength moderated the effects of standing balance on physical function. Accordingly, greater methodologic sophistication involving the use of electromyography and full-body motion capture may be required in future studies to evaluate our interpretations. Third, we only studied bipedal standing on a stable surface, and it would be informative to evaluate other balance tasks, for example, single extremity standing or standing on a perturbing surface, given their abilities to provide greater challenge to balance control.

In conclusion, our study demonstrated that in patients with symptomatic knee OA, knee extensor strength may moderate the standing balance–function association such that after adjusting for demographic and knee pain variables, standing COP displacement was positively related to physical function among patients with low knee strength but tended to be negatively related to physical function among patients with higher knee strength. These interactions have important research and clinical ramifications, but they call for further study.

AUTHOR CONTRIBUTIONS

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

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Pua had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Pua, Liang, Ong, Bryant, Lo, Clark.

Acquisition of data. Pua, Ong, Clark.

Analysis and interpretation of data. Pua, Liang, Ong, Clark.

Acknowledgements

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

We thank our patients for donating their time. We also thank Ms Tan Bee Yee, Ms Felicia Seet, Mr. Hanniel Lim, Ms Chong Hwei Chi, Ms Jennifer Liaw, and Mr. William Yeo for supporting this study. Finally, we thank the orthopedic surgeons from the Singapore General Hospital for allowing us access to their patients.

REFERENCES

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