Hip muscle weakness in individuals with medial knee osteoarthritis




To compare the strength of the hip musculature in people with symptomatic medial knee osteoarthritis (OA) with asymptomatic controls.


Eighty-nine people with knee OA and 23 controls age >50 years were recruited from the community. The maximal isometric strength (torque relative to body mass) of the hip abductors, adductors, flexors, extensors, and internal and external rotators was evaluated using hand-held dynamometry or a customized force transducer apparatus. Univariate linear models with age and sex included as covariates were used to compare muscle strength between groups.


In people with knee OA, significant strength deficits were evident for all hip muscle groups evaluated (P < 0.05). Compared with controls, strength deficits ranged from 16% (hip extensors) to 27% (hip external rotators) after accounting for differences in sex and age between groups.


People with knee OA demonstrate significant weakness of the hip musculature compared with asymptomatic controls. It is not clear if hip muscle weakness precedes the onset of knee OA or occurs as a consequence of disease. Findings from this study support the inclusion of hip strengthening exercises in rehabilitation programs.


Knee osteoarthritis (OA) is a chronic problem affecting a significant proportion of older people (1) and is a major cause of pain and disability. Muscle weakness, particularly of the quadriceps, is one of the earliest clinical signs of knee OA and has long been recognized as a hallmark of the disease. In fact, muscle weakness may precede disease onset and play a role in knee OA pathogenesis (2). Since muscle strengthening improves pain and function in knee OA (3), strengthening exercise is widely recommended for the condition (4). However, most research has focused on the role of quadriceps strength in knee OA, with little attention given to other lower extremity musculature. A thorough understanding of the impact of knee OA on other muscle groups is required if optimal strengthening pro grams are to be developed and employed. Increasing research suggests that hip muscle weakness may be associated with knee OA, but little research has evaluated this to date.

The medial tibiofemoral compartment is the most common site of disease at the knee, presumably due to greater loads borne across this compartment relative to the lateral compartment, and research evaluating risk factors for medial knee OA progression has highlighted the potential importance of hip muscle strength. Chang et al (5) demonstrated that a greater internal hip abduction moment during gait protected against ipsilateral medial OA progression over a period of 18 months. Chang and colleagues postulated that hip abductor weakness may result in additional contralateral pelvic drop, shifting the center of mass toward the swing extremity, which therefore increases forces across the medial compartment of the stance extremity and hastens disease progression. Although plausible, little research has directly evaluated whether people with medial knee OA do, in fact, demonstrate weakness of the hip abductors, or any other hip muscle group, compared with asymptomatic controls.

This study aimed to compare strength of the hip abductor, adductor, internal rotator, external rotator, flexor and extensor muscles in people with symptomatic medial knee OA and an asymptomatic control group. It was hypothesized that participants with knee OA would be significantly weaker in all hip muscle groups compared with controls.

Subjects and Methods


Eighty-nine people with knee OA and 23 controls age >50 years were recruited from the community. Knee OA participants were enrolled in a clinical trial evaluating hip strengthening, and data reported in this study were collected during baseline assessment for the trial (6). Participants had OA in at least one knee, fulfilling the American College of Rheumatology classification criteria (7), and reported an average knee pain score of >3 on an 11-point scale (where 0 = no pain and 10 = worst pain ever). Additional inclusion criteria were medial knee pain, medial compartment osteophytes and/or medial joint space narrowing, and knee alignment of ≤182° on a standardized semiflexed posteroanterior radiograph (corresponding to a mechanical axis of ≤180°) (8). Exclusion criteria were absent/doubtful OA (Kellgren/Lawrence grade 0 and 1) (9), knee surgery or corticosteroid injection within 6 months, current or past (within 4 weeks) corticosteroid use, systemic inflammatory arthritic conditions, or currently participating in supervised lower extremity strengthening.

Control subjects were completely without any knee, hip, or back pain for at least the previous 6 months. Consistent with the OA group, the following exclusion criteria applied: 1) current or past (within 4 weeks) oral corticosteroid use, 2) systemic arthritic conditions, 3) history of hip or knee joint replacement or tibial osteotomy, 4) any other muscular, joint, or neurologic condition affecting lower extremity function, and 5) inability to ambulate without a gait aid. Control participants also rated their knee pain using a numerical rating scale (on an 11-point scale, where 0 = no pain and 10 = worst pain ever). Any control participant scoring >0 was excluded. Controls did not undergo radiographic evaluation to exclude OA due to financial and ethical constraints. Ethical approval was obtained from the University of Melbourne Human Research Ethics Committee. All participants provided written informed consent.

Muscle strength.

For OA subjects, the symptomatic knee was the study extremity, and in cases of bilateral eligible knees, only the most symptomatic extremity was tested. For controls, the study extremity was randomly selected. Isometric strength of hip flexion, internal rotation, external rotation, abduction, adduction, and extension (in that order) was measured. After a single submaximal trial to familiarize participants with the procedure, 3 trials of maximal effort (each of 5 seconds duration) were performed, separated by 15 seconds of rest. The maximum force output (N) from the 3 trials was converted to torque (Nm) by multiplying by the resistance lever arm (m), then normalized to express strength relative to body mass (Nm/kg). These strength tests have excellent test–retest reliability in our laboratory (intraclass correlation coefficient 0.84–0.98).

Maximal isometric hip abduction and adduction were measured in the supine position with pelvis and contralateral extremity stabilization (10). A hand-held dynamometer (HHD; Nicholas Manual Muscle Tester, Lafayette Instruments) was placed over the lateral and medial femoral aspects of the distal thigh for abduction and adduction, respectively. Hip flexion, internal rotation, and external rotation were measured in a sitting position with hips and knees flexed to 90° (10). The HHD was placed immediately proximal to the superior pole of the patella for hip flexion and immediately proximal to the malleoli for rotation.

Hip extension was measured with the participant well stabilized in supine and the test extremity raised off the plinth and supported by a padded cuff (10). A force transducer connected to an electronic inclinometer was suspended from the ceiling using a chain, and the transducer-inclinometer device was attached to the cuff. The chain angle was ∼70° from the horizontal and the test hip was flexed ∼20° such that the test extremity was perpendicular to the force transducer (i.e., always summed to 90°). After correction for the extremity's gravitational weight, participants performed isometric hip extension while force data were recorded from the force transducer.

OA-related knee pain and physical dysfunction.

In the OA group, self-reported knee pain and difficulty with physical function were measured using the Western Ontario and McMaster Universities Osteoarthritis Index, where higher scores indicate worse symptoms. Pain scores ranged from 0–20, and physical function scores ranged from 0–68.

Statistical analysis.

Data analyses were performed using SPSS statistical software, version 16 (SPSS), and data were checked for normality prior to analysis. Since most data were normally distributed, parametric tests were used. Independent t-tests and chi-square tests were used to compare descriptive characteristics between groups. To account for potential influences of age or sex, muscle strength was compared between groups using univariate linear models with age and sex as covariates. To determine if muscle strength was correlated with disease severity in the OA cohort, Spearman's rho correlations were performed between muscle strength and radiographic disease severity data. An alpha level of 0.05 was used for all analyses.


Participant characteristics are shown in Table 1. Groups were similar regarding most demographic features with the exception of age, where the OA group was slightly older. The OA group reported moderate levels of pain, stiffness, and difficulty with physical function, and the degree of disease severity was equally dispersed among the group. Unadjusted and adjusted mean strength scores for both groups are presented in Table 2. After adjustment for differences in age and sex, significant weakness was evident in knee OA participants across all muscle groups. The magnitude of strength deficit ranged from 16% (hip extensors) to 27% (hip external rotators). Generally, muscle strength was not correlated with radiographic disease severity, with the exception of hip abductor strength (r = −0.29, P = 0.007), where more severe disease was associated with weaker hip abductors.

Table 1. Descriptive characteristics of the study sample*
 Knee OA (n = 89)Controls (n = 23)P
  • *

    Values are the mean ± SD unless otherwise indicated. OA = osteoarthritis; NA = not assessed; K/L = Kellgren/Lawrence.

  • Measured using an 11-point numerical rating scale, where 0 = no pain and 10 = worst pain ever.

  • Measured using the Western Ontario and McMaster Universities Osteoarthritis Index (pain range 0–20, stiffness range 0–8, and physical functioning range 0–68).

Age, years64.6 ± 8.360.3 ± 6.50.03
Height, cm167 ± 9166 ± 80.55
Mass, kg78.2 ± 15.872.7 ± 14.70.13
Sex, no. (%)   
 Female43 (48)16 (70) 
 Male46 (52)7 (30)0.07
Knee pain severity   
 Over past week4 ± 20 ± 0< 0.001
 Over past 48 hours7 ± 3NA 
Knee stiffness3 ± 2NA 
Difficulty with physical function24 ± 11NA 
Radiographic OA severity, no. (%)   
 K/L grade 230 (34)  
 K/L grade 329 (32)  
 K/L grade 430 (34)  
Alignment, degrees176.8 ± 2.5NA 
Table 2. Comparison of hip muscle strength between knee OA and control participants
Torque, Nm/kgKnee OA (n = 89)Controls (n = 23)Adjusted mean difference, %*P
Mean ± SDAdjusted mean ± SEM*Mean ± SDAdjusted mean ± SEM*
  • *

    Strength scores adjusted for age and sex. OA = osteoarthritis.

Hip flexion0.77 ± 0.230.76 ± 0.031.00 ± 0.391.03 ± 0.0526< 0.001
Hip extension1.86 ± 0.631.82 ± 0.052.05 ± 0.602.17 ± 0.10160.002
Hip internal rotation0.41 ± 0.160.41 ± 0.020.49 ± 0.200.51 ± 0.03200.011
Hip external rotation0.37 ± 0.150.36 ± 0.010.46 ± 0.160.49 ± 0.0327< 0.001
Hip abduction0.86 ± 0.290.86 ± 0.031.13 ± 0.321.13 ± 0.0624< 0.001
Hip adduction0.76 ± 0.290.76 ± 0.031.00 ± 0.351.03 ± 0.0626< 0.001


This study compared hip muscle strength in people with symptomatic medial knee OA with asymptomatic individuals. It was hypothesized that the hip flexors, extensors, rotators, abductors, and adductors would be weaker in those with OA. Results confirmed this hypothesis, although it is not clear from our cross-sectional design whether hip muscle weakness precedes development of knee OA or is a consequence of disease.

Lower-extremity muscle weakness is a hallmark of knee OA and may play a role in disease pathogenesis. Several studies have shown that people with knee OA are 20–40% weaker in relative quadriceps strength than healthy controls (11). Little research has directly compared hip muscle strength between people with knee OA and asymptomatic controls. Although Yamada et al (12) reported reduced absolute hip abductor strength and increased adductor strength in people with knee OA compared with healthy controls, conclusions from this study must be viewed cautiously. Since muscle strength was reported as the absolute force applied to the transducer rather than as a torque, differences in extremity length and body mass between subjects were not accounted for. Furthermore, force measures were taken in positions where the weight of the lower extremity detracted from the recorded force, and the measurement device was applied at the ankle rather than the distal thigh, which may have allowed other muscle groups or knee joint pain/laxity to influence force scores. Finally, only mean data were reported, and statistical analyses comparing muscle force between groups were not performed, so it is not clear whether muscle strength was, in fact, different between people with knee OA and controls.

The temporal relationship between hip muscle weakness and knee OA is unclear. Some evidence suggests quadriceps muscle weakness may precede knee OA development (2) and, although no study has evaluated hip muscle strength in relation to disease incidence, it is possible that hip muscle weakness may precede disease onset. Indirect evidence supporting a pathogenetic role for hip abductor weakness comes from a study of disease progression. In an 18-month study, every additional unit of normalized internal hip abductor moment during gait was associated with a 43% reduction in the risk of ipsilateral knee OA progression (5). Also, Mundermann et al (13) reported higher peak internal hip abduction moments during gait in mild to moderate knee OA compared with severe disease, further suggesting a possible protective effect of hip abduction strength against progression. Although both studies attributed internal hip abduction moments to hip abductor strength, this was not directly measured.

Conversely, hip muscle weakness may be a consequence of knee OA. Inhibition of muscle activation and muscle fiber atrophy are important mechanisms underlying quadriceps weakness (11), and it is possible that similar mechanisms may be involved at the hip. Given that people with knee OA frequently adopt compensatory gait patterns in response to chronic pain and structural pathology, it is possible that hip muscle weakness may result from using such strategies. For example, increased lateral trunk lean is often employed by patients to offload the medial knee (14). Although this may reduce medial load during gait, it would also reduce demand on the hip abductor muscles, potentially leading to hip abductor weakness over time. Finally, it is also quite probable that generalized weakness of all hip muscle groups results from relative disuse due to the reduced activity levels associated with knee OA.

Findings from our study have important clinical implications. Our data support the inclusion of hip muscle strengthening into rehabilitation programs for knee OA. Results from our recent clinical trial evaluating the efficacy of hip strengthening for knee OA (6) demonstrate improved pain and physical function when the hip abductors and adductors are strengthened, but no change in knee joint loading. Since quadriceps strengthening has traditionally formed the cornerstone of exercise therapy for knee OA, it is possible that patient outcomes may be optimized if hip strengthening is combined with quadriceps strengthening. Future research should evaluate this combination, as well as determine which patient subgroups may best benefit from hip strengthening. Our results suggest that hip strengthening is warranted for most muscle groups irrespective of radiographic disease severity. However, it is likely that hip abductor strengthening, in particular, is most needed in patients with severe disease given that this muscle group was the only one to exhibit a significant inverse correlation between strength and disease severity.

There are some limitations to our study. Our control group was slightly younger and comprised of a greater proportion of women than our OA group. However, we accounted statistically for these differences, and significant strength differences in all hip muscle groups persisted. Because of the lack of radiographs (due to ethical and financial constraints), we cannot conclusively rule out the presence of radiographic knee OA in our control group; however, we did ensure these people had no knee symptoms or any history of injury or trauma likely to predispose to knee OA. Similarly, we cannot exclude the possibility of hip OA coexisting in our knee OA cohort since we did not obtain hip radiographs. Although the presence of hip OA in some or all of our knee OA participants may explain the hip muscle weakness we observed, this is unlikely since participants were asymptomatic at the hip and underwent the same screening as control participants with respect to hip pathology, yet still demonstrated relative weakness.

In summary, this study compared hip muscle strength in people with symptomatic medial knee OA and asymptomatic controls. Significant weakness of the hip flexors, extensors, internal and external rotators, abductors, and adductors was observed in those with knee OA. Findings support the inclusion of hip strengthening exercises into rehabilitation regimens for knee OA.


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 submitted for publication. Dr. Hinman 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. Hinman, Hunt, Wrigley, Bennell.

Acquisition of data. Hunt, Creaby, McManus.

Analysis and interpretation of data. Hinman, Hunt, Wrigley, Bennell.