Research Article

*Correspondence to Leena Sharma, Division of Rheumatology, Feinberg School of Medicine, Northwestern University, 240 East Huron, Suite 2300, Chicago, IL 60611

Funded by:
  NIH; Grant Number: P60-AR-48098, R01-48748, R01-AR-46225, RR-00048

Received: 25 March 2005; Accepted: 9 August 2005

10.1002/art.21406  About DOI

In knee osteoarthritis (OA), the medial tibiofemoral compartment is the most common site of disease. The susceptibility of the medial compartment to OA development may relate to greater load distribution (i.e., 60-80%) to the medial than the lateral compartment, even in healthy knees, during gait. Excessive medial compartment loading is widely believed to contribute to medial OA progression. Because direct measurement of knee load is invasive, external knee adduction moment during gait, a correlate of medial load, has been used in knee OA studies ([1]). In keeping with the concept that load influences progression, a greater knee adduction moment predicted a greater likelihood of medial OA progression ([2][3]).

In theory, reduction of medial load may have a beneficial disease-modifying effect on medial knee OA, i.e., it may slow the rate of OA progression. However, it is unclear how to achieve this reduction. Altering certain kinematic or kinetic parameters during gait could, theoretically, reduce medial load. Whether any such parameters protect against knee OA progression has not previously been examined. Such information would guide and direct the development of novel rehabilitative interventions to delay medial knee OA progression.

A potentially protective kinetic parameter is the internal hip abduction moment. During the single-limb stance phase of gait, weakness or decreased torque generation of the hip abductor muscles in the stance limb causes excessive pelvic drop in the contralateral swing limb ([4]). This drop shifts the body's center of mass toward the swing limb, thereby increasing forces across the medial tibiofemoral compartment cartilage of the stance limb (Figure 1). The magnitude of hip abductor muscle torque generation during walking, the most common human weight-bearing activity, can be captured in quantitative gait analysis as the internal hip abduction moment.

Based on the proposed mechanism, a greater internal hip abduction moment during gait may prevent excessive medial compartment loading and potentially protect against ipsilateral medial OA progression. We tested the hypothesis that a greater peak internal hip abduction moment assessed during quantitative gait analysis is associated with a reduced likelihood of ipsilateral medial tibiofemoral OA progression, and we examined whether any effect persists after adjusting for potential confounders.

PATIENTS AND METHODS

Participants.

Mechanical Factors in Arthritis of the Knee (MAK) is a natural history study of knee OA at Northwestern University. As previously described ([5]), MAK participants were recruited from several community sources. Inclusion and exclusion criteria were based on findings of a National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS)/National Institute on Aging (NIA) workshop ([6]). Inclusion criteria were definite osteophytes in one or both knees and at least a little difficulty (graded using a Likert scale) with 2 or more items in the Western Ontario and McMaster Universities Osteoarthritis Index physical function scale. Exclusion criteria were a corticosteroid injection within the previous 3 months, avascular necrosis, inflammatory arthritis, periarticular fracture, Paget's disease, villonodular synovitis, joint infection, ochronosis, neuropathic arthropathy, acromegaly, hemochromatosis, Wilson's disease, osteochondromatosis, gout, pseudogout, osteopetrosis, and bilateral total knee replacement or plans for replacement within the next year.

Additional exclusion criteria for the current study were replacement of any joint in a lower extremity, lateral tibiofemoral OA bilaterally (based upon the presence of grade 2 definite narrowing, or worse, using the 0-3 scale of the Osteoarthritis Research Society International [OARSI] atlas). The sample in this study included all MAK participants who had quantitative gait analysis performed. All of these participants returned for followup. All participants provided informed consent. Institutional Review Board approval was obtained.

Quantitative gait analysis to measure hip joint moments.

Fifty-seven MAK participants underwent quantitative gait analysis (at Rush-Presbyterian-St. Luke's Medical Center) within 1 month of their MAK evaluation. The Computerized Functional Testing Corporation (Chicago, IL) system was used, including 4 optoelectronic cameras (Qualisys, Gothenburg, Sweden) with a sampling frequency of 120 Hz and a single multicomponent force plate (Bertec, Columbus, OH). Six passive markers were used in the link segment model. They were placed at the following bony landmarks: the lateral-most aspect of the superior iliac crest, the greater trochanter, the lateral joint axis line of the knee, the lateral malleolus, the lateral aspect of the calcaneus, and the base of the fifth metatarsal. Inverse dynamics were used to calculate the external moments (in 3 planes) at the hip, knee, and ankle joint centers, using the 3-dimensional kinematic data acquired with the cameras along with the ground reaction forces and moments obtained with the force plate.

The external moments were calculated by taking into account the moment about the joint center (created by the ground reaction force) and inertial forces. An external moment is equal and opposite to the internal moment created by muscles, soft tissues, and joint contact forces. The between-session reliability of hip abduction moment measurement for this laboratory was studied in 10 persons. The intraclass correlation coefficient (model 2.1) was 0.83. The examiner and the investigator (DH) processing the gait data were blinded to all radiographic data.

Measurement of covariates.

Knee pain severity was measured using 100-mm visual analog scales, with separate scales for the right and left knees. The scales were anchored at 0 =  no pain  and at 100 =  pain as bad as it could be, and  standardized instructions for assessment were given.

Physical activity was assessed using the Physical Activity Scale for the Elderly (PASE), a self-report measure of global activity including recreational, occupational, and household activities (higher score indicates greater activity) ([7]). PASE was designed to assess activities commonly engaged in by older persons, and its construct validity and test-retest reliability have been demonstrated in community-dwelling older adults ([7]).

Using knee radiographs acquired with the protocol described below, disease severity was assessed using the Kellgren/Lawrence (K/L) scale. In the K/L grading system, 0 = normal, 1 = possible osteophytes, 2 = definite osteophytes, possible joint space narrowing, 3 = moderate osteophytes, definite narrowing, some sclerosis, possible attrition, and 4 = large osteophytes, marked narrowing, severe sclerosis, definite attrition.

To assess the severity of varus malalignment, an anteroposterior full-limb radiograph was obtained according to a previously described protocol ([5]). Alignment was measured by one reader (LS) as the angle formed by the intersection of the lines connecting the centers of the femoral head and intercondylar notch with the lines connecting the centers of the ankle talus surface and tibial spine tips. Our reliability with this approach is high, as previously reported ([5]).

Hip symptom presence was defined as pain, aching, or stiffness lasting at least 1 month during the previous 12 months. Hip OA presence was defined using the American College of Rheumatology (ACR) clinical criteria ([8]).

Assessment of OA progression.

Bilateral radiographs of the knees with weight bearing were obtained at baseline and 18 months, following the Buckland-Wright protocol ([9]). This protocol meets criteria set by the NIAMS/NIA workshop ([6]) and the OARSI ([10]). Knee position, beam alignment, magnification correction, and measurement landmarks were specified. The semiflexed position superimposed the anterior and posterior medial tibial margins. Tibial rim alignment and tibial spine centering in the notch were confirmed fluoroscopically before radiographs were obtained. Radiographs were obtained in 1 unit by 2 trained technicians. Foot maps made at baseline were used at followup.

Medial tibiofemoral OA progression was defined as any worsening in the radiographic medial joint space grade between baseline and 18 months. Illustrated OARSI atlas grades (none, possibly, definitely, or severely narrowed joint space) ([11]) were used by 1 reader (LS). Reliability of radiographic grading (of joint space and K/L scoring) by the reader was very good (  = 0.85-0.86). The knee radiograph reader was blinded to the gait analysis data.

Statistical analysis.

Knees with a tibiofemoral joint space grade of 3 at baseline were excluded from analysis, since further progression was not possible. In descriptive analyses, hip abductor moment means were estimated from ordinary least squares regression. Since the rationale for our hypothesis did not support analysis of only 1 knee, we included both knees, using generalized estimating equations (GEE), which validly allow data from both knees to be included in the analysis ([12]). Logistic regression using GEE was used to assess the effect of increased hip abduction moment on the odds of OA progression. Results are presented as odds ratios (ORs) per unit of hip abduction moment. An OR of <1 represents a protective effect against OA progression; an associated 95% confidence interval (95% CI) that excludes 1 denotes a statistically significant effect. Hip moment data were normalized for body weight and height. Analyses were adjusted for age (treated as a continuous variable), sex, gait speed (continuous), knee pain (continuous), disease severity (as reflected by K/L score coded as indicator variables using K/L = 0 as the reference), severity of varus malalignment (continuous, with varus as a positive value, neutral as 0, and valgus as negative), physical activity (continuous), hip symptoms (presence/absence), and hip OA (presence/absence).

RESULTS

Our sample consisted of 57 participants (63% women) contributing 103 knees at risk for OA progression. The mean ± SD age was 67 ± 8.7 years and the mean ± SD body mass index was 29 ± 4.1 kg/m². The majority (72%) of the 103 knees had mild OA, without definite joint space narrowing. The remaining 28% of the knees had definite (but not severe) medial narrowing.

Figure 2 shows the trajectories of the internal hip abduction moment for a knee that progressed and a knee that did not progress from baseline to 18 months. The peak hip abduction moment occurs at the early stance phase of gait. The nonprogressing knee had a greater peak hip abduction moment than the progressing knee.

As shown in Table 1, the mean ± SD peak internal hip abduction moment in all knees was 4.41 ± 0.11 (% body weight × height). The peak internal hip abduction moment was greater in knees that did not progress than in knees that did progress over 18 months.

Next, we examined the relationship between the internal hip abduction moment at baseline and the likelihood of medial OA progression from baseline to 18 months (Table 2). Greater internal hip abduction moment had a protective effect, i.e., lowered the odds of progression, and this effect persisted after adjusting for potential confounders, including age, sex, gait speed, knee pain severity, physical activity, knee OA severity (or, in alternate models, severity of varus malalignment), hip symptoms, and hip OA presence. The hip abduction moment/progression relationship persisted after further adjustment for the external knee adduction moment (adjusted OR 0.27, 95% CI 0.11-0.69).

DISCUSSION

A greater internal hip abduction moment measured during quantitative gait analysis at baseline protected against medial tibiofemoral OA progression during the following 18 months. The odds of medial OA progression were reduced by 50% with an additional 1 unit of hip abduction moment. This strong protective effect persisted after adjusting for potential confounders.

In the study of knee OA, frontal-plane knee joint mechanics, such as varus or valgus alignment and laxity, as well as varus or valgus forces and moments acting on the knee, have received attention ([5][13-16]). However, the knee joint does not function in isolation from the rest of the lower limb kinematic chain during weight-bearing activities; hip and ankle/foot mechanics may influence knee joint load during gait.

The relationship between hip and knee mechanics in knee OA gait has been examined by McGibbon and Krebs ([17]). They focused on sagittal plane alterations in hip, knee, and ankle mechanics that develop presumably to compensate for ipsilateral knee OA disease. In their study, persons with OA had reduced knee extension concentric power and increased hip extension eccentric power when compared with healthy age-matched elderly subjects. They proposed that the changes in knee and hip power might be a compensation mechanism to avoid using the quadriceps, and thereby reduce articular loads. Their study illustrates the importance of considering the whole kinematic chain to better understand knee loads in individuals with knee OA. Because their focus differed from ours, they did not describe frontal-plane hip joint mechanics.

The contribution of frontal-plane hip joint mechanics to knee joint loading has not been examined. Especially since few knee muscles specifically provide knee frontal-plane stability, hip frontal-plane muscles may play an important role in regulating medial/lateral knee load distribution and providing frontal stability. In addition, hip muscles have a large cross-sectional area and can generate significant forces to control load. Fredericson and colleagues ([18]) found decreased hip abductor strength (measured with a hand-held dynamometer) in runners with iliotibial band syndrome (ITBS) when compared with the healthy limb and with runners without ITBS. After 6 weeks of hip abductor strengthening, these runners had increased hip abductor muscle strength and returned to pain-free running. It is plausible that decreased hip abductor activity might lead to excessive tensile stress on the lateral knee structures, such as the iliotibial band, and increased load on the medial compartment. The results of our study provide some longitudinal evidence for this theoretical framework. In order to capture the magnitude of the hip abduction muscle torque generated during a common weight-bearing activity, we focused on the hip abduction moment.

To our knowledge, this is the first longitudinal study on the effect of the hip abductor moment during gait on the course of medial knee OA. While surface electromyogram (EMG) can be performed during quantitative gait analysis, needle EMG is a superior, albeit more invasive, approach. However, EMG is classically used to assess motor unit recruitment and the temporal characteristics of muscle contraction, which are not well correlated with torque generation.

The major source of hip abduction moment magnitude is hip muscle strength. The hip joint ligaments and capsule also make a small contribution; there is no established way to account for this contribution. Hip symptoms could inhibit hip abductor muscle torque generation during gait ([19]) and could reflect an OA process in the hip that in and of itself contributes to knee OA progression. Only 5 participants met the ACR criteria for hip OA. We adjusted for both the presence of hip OA and the presence of hip pain without any impact on the protective effect of the hip abduction moment magnitude.

We also considered other potential confounders. Slower walking speed is associated with lower hip and knee joint moments ([20][21]) and may accompany progressive knee OA. Greater physical activity could theoretically be associated with greater hip torques and protection against OA progression. Greater varus malalignment or knee adduction moment each increases the likelihood of medial knee OA progression, and, in theory, could be linked to a reduction in hip abductor muscle torque. A strong protective effect of internal hip abduction moment persisted after adjusting for these factors.

As is almost always the case in knee OA cohorts in the US, the study sample was, on average, overweight. Because quantitative gait analysis uses skin markers, skin movement is an inherent limitation in any gait study of knee OA. One way to address this concern, placement of markers in bone, is of course not feasible. Measurement muddiness introduced by skin movement, if anything, would reduce the likelihood of identifying a relationship. Despite this, we were able to detect a protective effect.

The magnitude of increase in hip abduction moment with hip exercise is not known, and would be an important focus of an interventional study. Our results suggest the need for future studies to examine the therapeutic effect of interventions targeting hip abductors. Such interventions may be disease modifying, either on their own or as an adjunct to pharmacologic therapy.

In conclusion, a greater hip abduction moment during gait at baseline protected against ipsilateral medial OA progression from baseline to 18 months. The likelihood of medial tibiofemoral OA progression was reduced 50% per 1 unit of hip abduction moment.