Strength

Muscle weakness is a well-accepted impairment in knee OA (see Bennell and colleagues for an in-depth review). However, assessment of strength is not straightforward and has not always been well conducted in OA studies. Clinical methods may not be as sophisticated as research methods, but should follow similar principles. The ability to generate muscle force is a function of muscle cross-sectional area and the ability to recruit and fire descending alpha motor neurons to muscle fibers at sufficient frequencies. That force then acts in conjunction with the muscle’s moment arm to generate a torque or moment about a particular joint.

Ideally, strength should be measured to reflect this muscle-generated torque. Torque is thus measured as the product of the force (in Newtons) exerted at the point of attachment of a force-measuring transducer to the limb, and the distance (in meters) of that attachment from the axis of rotation of the joint in question; thus strength is reported in units of Newton meters (N m). Strength varies with body size (which affects muscle size and the moment arm length), in patients with OA as in athletes. In some studies on OA, the investigators have failed to normalize the torque measurements for differences in body size (most logically body mass in kilograms).

Results of studies that have correctly measured strength as torque, and normalized for body mass differences (ie, N m/kg) show that patients with knee OA are 20% to 40% weaker in relative quadriceps strength than healthy controls. Although other lower-limb muscles in knee OA have received less attention, strength of the hip muscles also seems to be reduced with 1 study reporting strength deficits ranging from 16% (hip extensors) to 27% (hip external rotators) compared with controls.

There are several contributors to muscle weakness in individuals with knee OA. Pain, anxiety, motivation, effusion, muscle atrophy, and aberrant joint mechanics can all contribute to a loss of measurable strength. Some of the weakness in relation to body size seen in OA is likely caused by the obesity that is also commonly present, as the proportion of total body mass made up of force-generating muscle is by definition reduced. Primary deficits in muscle strength may be associated with muscle fiber atrophy (ie, loss of muscle cross-sectional area), reduced ability to activate muscle fibers, or both.

There are few studies investigating muscle atrophy in OA. Ikeda and colleagues found that quadriceps cross-sectional area was significantly reduced by an average of 12% in women with incident radiological OA without symptoms, compared with women matched for age and body mass with no signs of OA. In later stage disease, more obvious signs of actual muscle fiber atrophy have been reported. In patients before knee replacement, Pettersen and colleagues found quadriceps lean muscle cross-sectional area to be 12% lower in the affected limb compared with the contralateral lateral side. Thus, it seems likely that at least some of any strength loss in OA is caused by loss of muscle cross-sectional area.

When muscle atrophy cannot explain the full extent of muscle weakness, inhibition in the ability to activate muscle is implicated. Detection of inhibition is primarily a research tool, whereby an electrical stimulus over the muscle is applied during a maximal contraction to determine whether the patient’s voluntary activation is less than maximal. Overall, there is a large variation in the results of studies assessing the extent of maximal voluntary muscle activation typically possible in OA. For example, 1 study found that patients before high tibial osteotomy had an average quadriceps activation of 71% (ie, 29% inhibited), with a range of 41% to 86%. Pettersen and colleagues found that inhibition explained most of the strength variance in end-stage knee OA. Despite some variation in research findings, the literature overall does show that patients with OA commonly exhibit impaired activation compared with healthy controls, as was the conclusion of a recent meta-analysis of published studies. This analysis also concluded that there was some evidence for activation deficits in the unaffected opposite limb.

The causes of muscle inhibition in knee OA are likely to be multifactorial. Pain is commonly presumed to be a major source of inhibition in the ability to voluntarily activate muscle surrounding arthritic joints. Pain relief via local anesthetic injection resulted in an 11% to 12% increase in percentage activation of the quadriceps in people with knee OA, confirming that at least some of the reduced voluntary activation observed in knee OA is caused by the presence of pain inhibition. Joint effusion on its own has also been found to be a potent inhibitor of maximal muscle activation in non-OA studies, even at low levels that might otherwise not be deemed clinically important (see Wrigley for review of classic work in this area). Effusion is often associated with OA and associated pain. Experimentally induced knee joint effusion has previously been found to generate pathologic gait changes. Recently, Rutherford and colleagues found that 17 medial patients with OA who presented with clinical knee joint effusions (bulge test) actually showed increased quadriceps activation and hamstring duration during gait compared with 18 patients without effusion. The patients also demonstrated greater stance flexion and a reduced external knee extensor moment. There were no differences in walking speed or pain between the groups. This supports the notion that effusions should be addressed in patients with OA, as they may be a barrier to rehabilitation and function.

Knowledge of the sources of muscle weakness in a patient is important, as it will determine how restoration of muscle strength might be approached therapeutically. If the deficit is primarily due to atrophy, then a pure muscle-strengthening approach should be taken. Alternatively, if the deficit is primarily in the ability to activate an essentially normal muscle, then attention might be directed toward removing the inhibitory sources that prevent sufficient activation (such as pain and effusion), and retraining the patient to activate their muscles fully.

Muscle Activation Patterns

In recent years, there has been increasing interest in the patterns of muscle activation associated with knee OA. As muscles are often the major contributors to joint loading, their inefficient activation during walking and other functional tasks could be implicated in disease progression. It seems that some patients with OA are able to activate their knee muscles in a way that is most efficient for dealing with the commonly increased joint loading, especially of the medial compartment during gait, while satisfying the other requirements for weight support and propulsion. However, other patients adopt a less efficient strategy that involves activating many muscles in a less specific fashion that may increase overall joint loading.

Several cross-sectional studies have noted differences in muscle activation patterns comparing people with knee OA of varying disease severity to age-matched controls. Using surface electromyography (EMG), Hubley-Kozey and colleagues found that patients with moderate OA biased the activation of quadriceps and hamstrings toward the lateral components of these muscle groups in early stance, compared with healthy controls. Schmitt and Rudolph found that patients with mild to severe medial tibiofemoral joint OA showed increased activation of lateral hamstrings and gastrocnemius before foot contact and increased cocontraction of the lateral quadriceps and gastrocnemius as well as medial quadriceps and hamstrings during weight acceptance.

Astephen and colleagues found that patients with severe tibiofemoral OA could be distinguished from healthy controls by increased medial and lateral hamstring activity during stance; they were also different from those with moderate OA in terms of higher medial gastrocnemius activity in swing and early stance phases, but lower activity during late stance. In patients with a range of OA severity, Childs and colleagues found that muscles were activated for longer durations in the period just before and during stance compared with healthy age-matched and gender-matched controls. Heiden and colleagues found that cocontraction biased more toward lateral knee muscles increased with higher external knee adduction moments; this suggests that this pattern of muscle activation is directly linked to resisting associated increases in medial compartment loading. Thus, it seems that, as the disease progresses, a less specific and efficient activation of a greater number of knee muscles may become more apparent.

All the alterations in activation patterns in knee OA described in the literature are not necessarily consistent, and part of the variability may relate to methodological issues particularly in EMG measurement and analysis. Nevertheless, it seems that muscle activation patterns do differ in people with knee OA and may be influenced by the stage of OA disease and aspects of the disease process. For example, a recent study found that, in people with knee OA, those with a joint effusion had greater overall quadriceps activation and prolonged hamstring activation during midstance compared with those without a joint effusion. Whether these differences reflect the increasing difficulty in activating muscles efficiently for ambulation in the face of pain, effusion, and deteriorating joint mechanics, and/or are involved in disease progression itself, needs to be clarified. There is also little evidence that such patterns can be altered; nor is it clear what would constitute a desirable alteration.

Proprioception

Knee joint proprioception is important for the coordinated activity of surrounding muscles to protect against excessive movements, stabilize during static postures, and coordinate movement. Proprioceptive afferent information from mechanoreceptors, particularly in muscles but also in ligaments, capsule, menisci, and skin, contributes at the spinal level to arthrokinetic and muscular reflexes, which play a large part in dynamic joint stability. The information is also conveyed to supraspinal centers where it is integral to motor learning and the ongoing programming of complex movements. Abnormal proprioception could predispose to musculoskeletal lesions by altering the control of movement leading to abnormal stresses on tissues. Alternatively, a pathologic condition, effusion, and pain may impair proprioceptive information, possibly further compounding functional deficits.

Proprioception is typically measured in OA studies as conscious perception of joint position sense using accuracy of reproduction or threshold of movement detection tests. However, proprioception during normal function rarely requires conscious perception. Furthermore, just because a patient perceives proprioceptive information (consciously or unconsciously) does not mean that it is efficiently used by the nervous system to modulate movement. Measurements of knee position sense and knee motion sense are not well correlated and as such seem to indicate different aspects of knee proprioception and probably stimulate different receptors.

Most studies have found deficits in knee joint proprioception in patients with knee OA compared with similarly aged asymptomatic individuals. A recent study also showed consistent differences in knee proprioception between groups with and without knee OA across all knee movement directions (varus, valgus, flexion, and extension) suggesting a global, rather than a direction-specific, reduction in sensation in patients with knee OA. In addition to impaired proprioception at the affected knee joint, there is also evidence of proprioceptive deficits at other nonaffected sites including the contralateral knee and the elbow. This has led to suggestions that OA may be associated with a generalized defect in proprioception. The topic of proprioception in knee OA has recently been reviewed by Knoop and colleagues.

Disease Onset

There is some evidence to suggest that quadriceps weakness precedes the onset of knee OA and hence could increase the risk of disease development, particularly in women. The first longitudinal study to investigate this issue more than a decade ago demonstrated 15% to 18% lower baseline isokinetic quadriceps strength (adjusted for body and muscle mass) in women who went on to develop incident radiographic knee OA than in women who did not develop OA. This was not seen in men nor was there a relationship between hamstring strength and OA onset in either sex. Although the study had some methodological issues, including lack of adjusting for potential covariates, these results were subsequently confirmed in a larger longitudinal cohort study involving 3081 adults. In this study, higher isokinetic quadriceps strength (relative to body mass) was associated with a 55% to 64% reduced risk of developing knee or hip OA (self-reported) in women with similar, albeit nonsignificant, results in men (∼25% less risk). Support for a link between quadriceps strength and disease onset also comes from a study on a group of 94 people aged 35 to 54 years with chronic knee pain but normal radiographs at baseline, in which the maximal number of one-legged rises, a functional measure that can indicate quadriceps strength, predicted radiographic knee OA 5 years later.

More recent data from the large Multicenter Osteoarthritis Study (MOST) published in several papers have provided further insights into the relationship between muscle strength and OA development. In 1 study, absolute isokinetic knee extensor and flexor strength was adjusted for age, body mass index, hip bone density, surgical history, pain, and physical activity score in multivariate models, and related to the onset of symptomatic and radiographic OA 30 months later. Compared with the lowest tertile, the highest tertile of adjusted isokinetic knee extensor strength protected against development of incident symptomatic tibiofemoral or patellofemoral OA in both sexes. However, neither adjusted knee extensor strength nor the hamstring to quadriceps ratio was predictive of incident radiographic tibiofemoral OA. This suggests that, in this cohort at least, quadriceps muscle weakness may play a greater role in OA symptom development than structural changes per se. These results were confirmed using knee extensor specific strength (knee extensor torque divided by total thigh muscle mass). This study also found that absolute total thigh muscle mass, as measured by dual energy x-ray absorptiometry, was not related to OA incidence. This might be interpreted as implicating other sources of weakness than reduced contractile material (such as the inhibition discussed earlier). However, it would be interesting to further explore adjustment for the expected effects of body size on a volumetric quantity such as absolute muscle mass.

People who have sustained a knee joint injury, particularly a tear of the anterior cruciate ligament (ACL) or meniscus, are at an increased risk of developing knee OA compared with those with uninjured knees. Whether subsequent muscle weakness further predisposes these individuals to developing knee OA has only been investigated in 1 longitudinal study to date. In this study, concentric quadriceps muscle function measured using isokinetic dynamometry (60°/s, 5-repetition total work divided by body mass) at 6, 12, and 24 months following patellar tendon ACL reconstruction was not associated with the development of radiographic or symptomatic radiographic knee OA 10 to 15 years later. Further research is needed to determine the role of muscle function in influencing knee joint health in this patient population.

The link between knee proprioception and OA disease onset has been investigated in 2 recent large longitudinal studies using data from MOST. Active knee joint position sense measured without weight bearing at baseline was not significantly associated with incident radiographic or symptomatic knee OA or with the new onset of frequent knee pain over a 30-month follow-up. Although it was hypothesized that sensorimotor factors may interact to mediate OA risk, the combination of high absolute isokinetic knee extensor strength and better knee joint position sense did not protect against development of knee OA. These findings suggest that proprioception, at least as assessed by knee joint position sense, does not predict the onset of either radiographic or symptomatic knee OA.

Disease Progression

Although it seems that quadriceps muscle strength is related to OA disease onset, the longitudinal evidence to suggest that stronger muscles can protect against OA progression in those with established disease is conflicting. Most studies used isokinetic dynamometry to assess concentric quadriceps strength (albeit at various speeds) and different methods and grading systems to assess disease progression.

The earliest study, published in 1999, found that the mean absolute quadriceps strength of women with progressive OA (defined as worsening of the Kellgren and Lawrence grade over 2.5 years) was about 9% lower than those with radiographically stable OA. However, this strength difference between the groups was not statistically significant, which could be due to the small number of participants (17 out of 82) exhibiting radiographic progression. Strength relative to body mass did not differ between progressors and nonprogressors. The use of nonstandardized conventional radiographs for evaluating progression could also reduce the potential of this study to detect structural progression. Nevertheless, in support of this nonsignificant finding, a larger more recent study involving 265 individuals and using the supposedly more sensitive technique of magnetic resonance imaging (MRI) to assess tibiofemoral cartilage loss over 30 months also failed to find an effect of relative isokinetic quadriceps strength on disease progression. Conversely, a study using data from MOST showed that women in the lowest tertile of relative isokinetic quadriceps strength (normalized for height and weight) had an increased risk of tibiofemoral joint space narrowing over 30 months (odds ratio 1.69, 95% confidence interval 1.26, 2.28) compared with women in the highest strength tertile. These results suggest that in women, but not men, quadriceps weakness may be a risk factor for structural deterioration in those who already have knee OA. A differential gender result also concurs with previous findings for incident symptomatic knee OA in MOST. Segal and Glass postulated that this gender difference may relate to the strength capacity of women being closer to a threshold for risk, whereas greater absolute strength in men provides greater reserve such that a loss of strength is insufficient to lead to greater risk of progression.

Muscle strength may have differential effects depending on the joint involved. Greater isokinetic quadriceps strength (60°/s ) relative to body mass has been found to protect against cartilage degeneration in the lateral aspect of the patellofemoral joint. The findings were supported by another study showing a similar trend in women but not men, although that study used absolute isokinetic strength (60°/s) not strength relative to body size (eg, body mass). Quadriceps weakness could alter patellar tracking and walking biomechanics, which could lead to disproportionate compressive forces across the patellofemoral joint.

It has been proposed that the local mechanical environment may also influence the relationship between strength and disease progression. A study by Sharma and colleagues found that greater absolute quadriceps peak torque (at 120°/s) at baseline increased the risk of tibiofemoral joint disease progression in individuals with malaligned knees (defined as >5° deviation from the mechanical axis) or high-laxity knees (defined as >6.75° varus valgus deviation), but not in those who had neutral alignment or low-laxity knees. The investigators suggested that the inability of malaligned knees to evenly distribute muscle forces could result in focal stress, and the increase in muscle contraction to stabilize lax knees could lead to higher joint reaction forces. The inferred clinical implication of these results, espoused in an editorial that accompanied this study, was that quadriceps strengthening, the cornerstone of exercise therapy in OA, might actually be detrimental in those people with knee OA and malalignment or laxity. However, a more recent study in 265 elderly individuals has not supported the results of Sharma and colleagues. In this study, there was no relationship between the relative isokinetic quadriceps strength (60°/s) at baseline and loss of medial tibiofemoral joint cartilage using MRI in those with or without knee malalignment. As longitudinal studies assess relationships and cannot directly assess causality, clinical trials are needed to definitively investigate the effects of strengthening exercise on structural outcomes at different joint compartments, in different sexes, and in the presence of different mechanical factors.

Although the quadriceps muscle has been the major focus, there has also been interest in the role of the hip abductors. It has been suggested that stronger hip abductors may be associated with a reduced risk of knee OA progression in the ipsilateral knee. In an 18-month study of 57 patients with mild to moderate unilateral knee OA, every additional unit of normalized internal hip abductor moment during gait was associated with a 43% reduction in the risk of ipsilateral medial knee OA progression. Although hip abductor strength was not measured, the investigators suggested that the internal hip abduction moment might be primarily related to abductor muscle strength. Mundermann and colleagues reported higher peak internal hip abduction moments during gait in individuals with mild to moderate knee OA compared with those with severe OA, further suggesting a protective effect of hip abduction strength against knee OA progression.

Some studies have used the external knee adduction moment during gait to examine the indirect effects of muscle strengthening on OA disease progression, by virtue of the moment’s association with both medial compartment loading and structural disease progression. Of the 6 studies, only 2 found an effect of exercise in reducing the knee adduction moment; both were uncontrolled pilot studies. In 1 study, an 8-week, supervised, lower-limb strength and neuromuscular control exercises program including functional strengthening was evaluated in 13 patients with early knee OA. The investigators found that the peak knee adduction moment during gait was not significantly altered at completion of the exercise program but there was a 14% reduction in the adduction moment measured during a one-leg rise. However, the clinical relevance of small changes in loading only during a one-leg rise task is unknown. The other study of 6 patients found that a 4-week strengthening program involving the hip and other lower-limb muscles led to an average 9% reduction in the peak knee adduction moment during walking even though strength gains were not found.

In contrast to these pilot studies, several larger RCTs and an uncontrolled study of 40 people have failed to find changes in the knee adduction moment with strengthening programs targeting the quadriceps and hip muscles, despite strength gains and symptomatic improvements. In 1 study, 107 participants were stratified according to knee malalignment (more varus or more neutral) and randomized into either a 12-week, supervised, home-based, quadriceps strengthening group or a control group with no intervention. Quadriceps strengthening did not significantly alter the adduction moment in people with more malaligned or more neutral knees. In another recent RCT, a 12-week hip muscle-strengthening program did not affect the knee adduction moment in 89 people with medial knee OA. These results were further confirmed by a study involving 54 women randomized into a 6-month high-intensity progressive resistance training program or sham exercise. In summary, the evidence currently suggests that muscle-strengthening exercise does not alter the knee adduction moment in people with knee OA.

Limited longitudinal data do not support a link between proprioception and progression of knee OA in those with established disease. From MOST, active joint position sense was not associated with radiographic worsening over 30 months in 2440 people with mild disease. It is not known whether proprioceptive deficits become more important risk factors in later stage disease when greater impairments in the sensorimotor system are found.

Although the role of muscle function on structural disease development and progression has received most of the attention in the literature, its effects on functional progression are less well known. Sharma and colleagues conducted a 3-year longitudinal cohort study investigating factors contributing to poor physical functioning in 257 patients with knee OA. They found that in addition to factors such as age, reduced absolute quadriceps and hamstrings strength and poor proprioceptive acuity (measured as joint position sense) increased the likelihood of poor physical functioning as measured by the time to perform 5 repetitions of rising and sitting in a chair. Similarly, in a cohort of more than 2000 individuals, those with worse proprioceptive acuity (measured as active non–weight-bearing joint position sense) at baseline had slightly greater worsening of WOMAC (Western Ontario and McMaster Universities) pain scores (0.47 on a 20-point scale) and physical function scores (by 1.5 points on a 0- to 68-point scale) compared with those with the best proprioceptive acuity (for pain P = .05; for physical function P = .02) over 30 months. Lastly, a longitudinal cohort study found that reductions in hip abductor strength were related to poorer functional outcomes more than 3 years in people with knee OA. Taken together, these findings suggest that impaired muscle function is related to greater pain and declines in physical in those with knee OA.

Muscle Strength

Extensive research indicates that muscle strength can be improved with an appropriately targeted strengthening program. A large systematic review assessed the effectiveness of isolated resistance training in people with knee OA. Fourteen RCTs were identified that measured parameters of muscle strength and compared the effects of isolated resistance training to a non-exercising control group. In general, muscle strength improved significantly with resistance training (mean improvement of 17.4%, range 10.5% decrease to 49.5% increase), with 9 out of 14 studies reporting significant strength gains. Relative effect sizes for strength outcomes ranged from -0.04 to 1.52, with an average of 0.38, indicating small to moderate effects of resistance training in this patient group. Whereas/although the most research has targeted strength training to the quadriceps muscle, strength gains have also been observed with targeted training for the hamstrings and hip musculature. Strength gains are apparent with both supervised clinic-based programs and home exercise regimes, although it is presently not clear which mode of delivery is superior for achieving strength gains as this has not been directly evaluated. The magnitude of strength gain achieved with resistance training varies according to the intensity of training (resistance applied as well as frequency), patient adherence and the specificity of training, which probably explains the variation in strength change observed in the systematic review by Lange and colleagues. More research is needed to establish dose-response relationships with respect to resistance training for people with knee OA.

There has been interest in whether modalities such as EMG biofeedback or whole-body vibration platforms can assist in augmenting strength gains achieved with exercise in people with knee OA. A recent systematic review to determine the magnitude of the treatment effect for EMG biofeedback on quadriceps strength compared with that of placebo and traditional exercise interventions in both healthy and pathologic populations found the strongest effect in knee OA populations. However, more definitive evidence is needed to confirm the benefits. With regards to vibration, there is limited research in knee OA with conflicting results. Segal and colleagues recently found that in women with risk factors for knee OA, the addition of vibration to a twice-weekly lower-limb exercise program did not result in significantly greater improvements in lower-limb strength or power than did performance of the exercise program without vibration. In another small RCT, knee muscle strength gains were significantly increased compared with control. A recent systematic review on whole-body vibration programs in older populations in general concluded that the effects do not seem to be greater than those achieved with conventional exercise alone.

Muscle Atrophy and Activation

Increases in muscle strength with resistance training are probably only partly explained by muscle hypertrophy, and research is required to investigate the relative contributions of neural adaptation and muscle hypertrophy to the strength increases observed.

One study evaluated the effects of isokinetic training on the cross-sectional area of the quadriceps and hamstrings in people with bilateral knee OA. Significant within-group increases in cross-sectional area were observed for both muscle groups with both a concentric and a combined concentric-eccentric training program.

Some studies have demonstrated improvements in voluntary activation of the quadriceps following an exercise regime, usually including quadriceps strengthening exercise. Mean within-group increases in activation ranging from 5 to 14% have been reported. Other treatment modalities that have been shown to influence motor neuron pool excitability in healthy subjects with experimentally effused knee joints include transcutaneous electrical nerve stimulation (TENS) and cryotherapy. Pietrosimone and colleagues randomly allocated participants with knee OA to receive either 45 minutes of TENS, 20 minutes of focal knee joint cooling with ice bags or control intervention (sitting quietly for 20 minutes) and evaluated the effect of treatment on the quadriceps central activation ratio. Application of TENS resulted in a significantly higher percent change in activation ratio scores compared with control at 20 min, 30 min, and 45 min. Focal knee joint cooling resulted in significantly higher percent change activation ratio scores compared with the control group at 20 min only. No significant differences in percent change for the central activation ratio were found between the TENS and the focal knee joint cooling group. The same investigators also showed that the addition of TENS to an exercise program increases quadriceps activation beyond that observed with exercise alone. These studies provide evidence that modalities such as TENS and cryotherapy may be useful for increasing muscle activation in people with knee OA, at least in the quadriceps.

Scopaz and colleagues evaluated whether pre-treatment magnitude of quadriceps activation predicted the change in quadriceps strength with exercise therapy in knee OA. The investigators hypothesized that people with lower magnitudes of muscle activation would have smaller gains in strength following exercise, when compared with those with higher magnitudes activation. Although correlations demonstrated that baseline quadriceps activation was positively associated with absolute isometric quadriceps strength at both baseline and 2-month follow-up, the level of relative quadriceps activation at baseline did not predict post-exercise absolute quadriceps strength when data were controlled for baseline absolute strength and type of exercise therapy. This data would suggest that factors other than baseline voluntary quadriceps activation may be more important in determining response to strengthening exercise. However, Pietrosimone and Saliba did recently show that changes in voluntary activation predicted 47% of the change in quadriceps strength in a group of patients with OA undergoing 4 weeks of supervised resistance training. Thus improvement in voluntary activation was linked to improvement in strength. This would suggest that deficits in activation could be investigated and targeted as part of OA rehabilitation.

Although impairments in muscle activation patterns during gait are associated with knee OA (see earlier), there currently seem to be no studies evaluating whether exercise can change this specific impairment in people with OA, probably because it is not clear what constitutes an optimal or desirable muscle activation pattern in these people. However a recent study has shown that application of a new therapy, stochastic resonance electrical stimulation combined with a neoprene knee sleeve, significantly decreased the ratio of vastus lateralis to lateral hamstring activity during walking. Similarly, a study in people with medial knee OA found that a valgus producing knee brace significantly reduced the cocontraction of both the vastus lateralis-lateral hamstrings and vastus medialis-medial hamstrings. Thus, it is possible to alter muscle activation patterns but whether this translates into clinical or structural benefits is not yet known.

Proprioception

Although many treatment strategies are postulated to improve proprioception, relatively few intervention studies in knee OA evaluate proprioception in their assessment of treatment efficacy. There is some evidence that impairment in proprioceptive acuity associated with knee OA may be enhanced with exercise training more so with exercise that specifically targets the proprioceptive system. One novel program involves a computer game, where seated participants control movement of a snake on the computer screen by stepping on 4 foot pedals in multiple directions at increasing speeds. In addition to training the proprioceptive system, task performance involves some resistance applied to the foot. This program has been evaluated alongside strength training and a no-exercise control in RCTs conducted by the same group of investigators, although it is not clear if the data reported in each paper were collected from 3 separate cohorts of people with knee OA or a single group. Within-group analyses demonstrated significant improvements in absolute error in knee angle repositioning from baseline to follow-up in the proprioceptive training group in all 3 studies (changes approximating 2–3° in magnitude). No change was evident within the no exercise control group or within the group undergoing seated concentric & eccentric quadriceps training. Another RCT, which evaluated the additive effect of a sensorimotor training program using sling suspension to routine physical therapy, found that the program resulted in improved knee repositioning compared with physical therapy without the sensorimotor component (mean change 1.9° ± 1.7). In contrast, another randomized study demonstrated that the addition of kinesthesia and balance exercises to a routine strengthening program did not offer any additional improvement in proprioceptive acuity than a strengthening program alone.

The effects of whole-body vibration on proprioceptive function have been investigated in knee OA. An RCT compared whole-body vibration on a stable platform to whole-body vibration on a balance board to a non-treated control group. Movement detection threshold was significantly improved in the vibrating balance board group compared with controls and there was a tendency for the vibrating stable platform group to perform better than the controls. It is not clear from this study whether improvement in proprioception was caused by/because of the whole-body vibration or simply an effect of the training on the balance board.

There is some evidence to suggest that knee bandaging or bracing can improve proprioception in knee OA, but findings are inconsistent across the literature. Neither taping, stimulating massage, magnetic knee wraps nor stochastic resonance electric stimulation seem to have an effect/affect on proprioception in people with knee OA. Thus it seems that some but not all treatments show promise for enhancing proprioceptive acuity in knee OA. However reported mean changes are generally quite small and it is unclear if such changes are clinically relevant or indeed, merely within the realm of measurement error.