Article Text
PDF
Abstract
Objective Varus and valgus alignment increase medial and lateral tibiofemoral load. Alignment was associated with tibiofemoral osteoarthritis progression in previous studies; an effect on incident osteoarthritis risk is less certain. This study tested whether alignment influences the risk of incident and progressive radiographic tibiofemoral osteoarthritis.
Methods In an observational, longitudinal study of the Multicenter Osteoarthritis Study cohort, full-limb x-rays to measure alignment were acquired at baseline and knee x-rays were acquired at baseline and knee x-rays at baseline and 30 months. Varus alignment was defined as ≤178° and valgus ≥182°. Using logistic regression and generalised estimating equations, the associations of baseline alignment and incident osteoarthritis at 30 months (in knees without baseline osteoarthritis) and alignment and osteoarthritis progression (in knees with osteoarthritis) were examined, adjusting. For age, gender, body mass index, injury, laxity and strength, with neutral knees as referent.
Results 2958 knees (1752 participants) were without osteoarthritis at baseline. Varus (adjusted OR 1.49, 95% CI 1.06 to 2.10) but not valgus alignment was associated with incident osteoarthritis. 1307 knees (950 participants) had osteoarthritis at baseline. Varus alignment was associated with a greater risk of medial osteoarthritis progression (adjusted OR 3.59, 95% CI 2.62 to 4.92) and a reduced risk of lateral progression, and valgus with a greater risk of lateral progression (adjusted OR 4.85, 95% CI 3.17 to 7.42) and a reduced risk of medial progression.
Conclusion Varus but not valgus alignment increased the risk of incident tibiofemoral osteoarthritis. In knees with osteoarthritis, varus and valgus alignment each increased the risk of progression in the biomechanically stressed compartment.
Statistics from Altmetric.com
The load-bearing axis of the lower limb can be represented by a line extending from the femoral head centre to the ankle joint centre. In a varus (bow-leg) knee, this line passes medial to the centre of the knee, increasing force across the medial tibiofemoral compartment. In a valgus (knock-knee) knee, the axis passes lateral to the knee centre, increasing force across the lateral compartment. Animal studies and human studies of complicated fractures provided some early evidence that alignment may influence the development and progression of knee osteoarthritis.1
In recent years, natural history studies of primary knee osteoarthritis have revealed a link between alignment and subsequent osteoarthritis progression.2,–,6 The effect of varus and valgus alignment on the risk of incident knee osteoarthritis is less certain.5 7 Alignment was found to have a stronger effect in knees with moderate tibiofemoral osteoarthritis than in knees with mild osteoarthritis, presumably relating to the greater vulnerability of more diseased knees to altered load distribution.8 In view of this, it seems likely that any alignment effect on the risk of incident knee osteoarthritis is smaller, and possibly more difficult to detect, than the effect on progression.
Most paradigms of knee osteoarthritis development and progression posit a central role for local mechanical factors acting within a systemic milieu. There is a particularly compelling biomechanical rationale to support a role for varus and valgus alignment. However, few longitudinal cohort studies have examined the alignment effect on the risk of incident knee osteoarthritis. To advance understanding of the pathways to knee osteoarthritis development and to inform the development of non-invasive prevention strategies, it is important to clarify the impact of alignment in knees without established osteoarthritis.
In a prospective ancillary study to the Multicenter Osteoarthritis Study (MOST), we tested the hypotheses:
Varus and valgus alignment increase the odds of incident radiographic tibiofemoral osteoarthritis, in knees without tibiofemoral osteoarthritis at baseline.
Varus alignment increases the odds of medial osteoarthritis progression, and valgus alignment increases the odds of lateral osteoarthritis progression, in knees with osteoarthritis at baseline.
Methods
Sample
MOST is an observational cohort study of incident and progressive knee osteoarthritis in 3026 community-dwelling men and women, aged 50–79 years. Participants were recruited using mass letter and brochure mailings and community outreach campaigns and were enrolled at Iowa City, Iowa, or Birmingham, Alabama. To be eligible for MOST, individuals were required to have symptomatic knee osteoarthritis or characteristics that placed them at increased risk of developing it during the study.9 10 Exclusion criteria were: bilateral total knee replacement or plan for this within the next year; inability to walk without the aid of another person or a walker; a serious health condition that would limit longitudinal study participation; ankylosing spondylitis, psoriatic arthritis, reactive arthritis, or rheumatoid arthritis; dialysis; cancer other than non-melanoma skin cancer; or a plan to move from the area within 3 years.
The study protocol was approved by the institutional review boards at each participating site.
Measurement of varus–valgus alignment and other factors at baseline
Alignment was assessed from full-limb radiographs, including hip and tibio-talar joints, acquired at baseline using a previously described protocol.2 Participants stood with the tibial tubercle facing forward. The x-ray beam was centred at the knee at a distance of 2.4 m. A setting of 100–300 mA/s and 80–90 kV was used, depending on limb size and tissue characteristics. One anteroposterior radiograph of both limbs was obtained. The full limb of tall participants was included by using a 51×14 inch graduated grid cassette (Iowa) and a system of overlapping cassettes and simultaneously exposed subimages forming a stitched image (Birmingham).
Alignment (the hip–knee–ankle angle) was measured as the angle at the intersection of the line connecting the femoral head and intercondylar notch centres with the line connecting the ankle talar surface centre and tibial interspinous sulcus base. Image analysis11 was completed in batches of 50 by one of three trained readers using a customised program (Surveyor 3; OAISYS, Kingston, Ontario, Canada) and blinded to all other data. Each batch was reviewed by a manager before transmission. In a reliability study of 200 full-limb pairs assessed by the three readers, the interreader and intrareader intraclass correlation coefficients for the hip–knee–ankle angle were 0.95 and 0.96, respectively.12 In analyses, varus alignment was defined as 178° or less, valgus as 182° or greater and neutral as 179–181°.
Concentric knee extensor strength was measured for each lower limb using a Cybex 350 isokinetic dynamometer (Avocent, Huntsville, Alabama, USA) at 60°/s.13 The average torque (Nm) of four maximum effort repetitions was analysed. Medial–lateral laxity (°) was measured using a protocol and device previously described,14 consisting of a bench and attached arc-shaped track, and providing thigh and ankle immobilisation, a stable knee flexion angle and fixed medial and lateral load. Weight (kg) without shoes or heavy clothes was measured on a balance beam scale and height without shoes using a stadiometer. Injury was defined as any knee injury severe enough to limit the ability to walk without a gait aid for at least 2 days.
Knee x-ray acquisition and assessment
At baseline and 30 months, knee radiographs were acquired using the posteroanterior ‘fixed-flexion’ weight-bearing protocol,15 in which knees are flexed to 20–30° and feet internally rotated 10° using a plexiglass positioning frame (SynaFlexer; Synarc, Inc, San Francisco, California, USA). The right and left knees were imaged together on 14×17 film with a 72 inch film-to-focus distance. Lateral weight-bearing films were also obtained, following a Framingham Osteoarthritis Study protocol.16
An experienced rheumatologist and musculoskeletal radiologist independently assessed each posteroanterior film for Kellgren and Lawrence (K/L) grade and each posteroanterior and lateral film for medial and lateral joint space narrowing grade. The readers were blinded to clinical data and knew the time sequence of the images.17 The joint space narrowing grade was scored (0–3) separately for the medial and lateral compartments using a modified version of the Osteoarthritis Research Society International (OARSI) scale.18 19 Previous studies3 revealed worsening of joint space narrowing over time not sufficient to move one full OARSI grade. When this occurred in knees with joint space narrowing at baseline, readers were instructed to use half grades, an approach that has been validated.19 If readers disagreed on whether incident osteoarthritis had developed or the joint space narrowing grade had worsened, the reading was adjudicated by a panel of three readers. Weighted kappas for agreement between the two readers were: K/L grade 0.79; medial and lateral joint space narrowing grade 0.81 and 0.86, respectively.
Definition of key outcomes
All outcomes were knee-based and assessed from baseline and 30-month radiographic images. Using the established and widely applied approach, radiographic knee osteoarthritis required definite osteophyte presence (K/L ≥2) at standard image size; incident osteoarthritis was defined as the new onset of K/L 2 or greater at 30 months, in knees graded K/L 0 or 1 at baseline.
Osteoarthritis progression was assessed in knees of K/L grade 2 or greater at baseline using a compartment-specific approach. Medial osteoarthritis progression was defined as any worsening of modified OARSI grade of medial joint space narrowing, and lateral osteoarthritis progression as any worsening of the lateral joint space narrowing grade. Knees with advanced osteoarthritis that could not progress further (K/L grade 4 or joint space narrowing grade 3) were excluded from analyses.
Statistical analysis
Both knees from each person were examined. Knee characteristics were calculated separately among knees without radiographic knee osteoarthritis (K/L <2) at baseline and knees with knee osteoarthritis (K/L ≥2) at baseline by alignment group (varus (≤178°), valgus (≥182°) and neutral (179–181°). Multiple logistic regression with generalised estimating equations, to account for potentially correlated observations for knees from the same person, was used to evaluate:
The relationship of varus and valgus alignment at baseline to incident knee osteoarthritis at 30 months, among knees without tibiofemoral osteoarthritis at baseline and therefore at risk of incident osteoarthritis.
The relationship of varus and valgus alignment to medial and lateral tibiofemoral osteoarthritis progression at 30 months, among knees with osteoarthritis at baseline and therefore at risk of progression.
Neutral knees constituted the reference group in all analyses. All analyses were adjusted for age (continuous), gender, body mass index (BMI) (continuous), knee injury (dichotomous), laxity (continuous) and extensor strength (continuous). Results from each model are reported as adjusted OR with associated 95% CI; a 95% CI excluding 1 represents a statistically significant association. In secondary analyses, the severity of varus and valgus alignment were analysed as continuous variables. Sensitivity analyses were run separately for men and women to determine if results were consistent across gender. Analyses were performed using SAS software version 9.2.
The funding source played no role in the study design and conduct, the collection, management, analysis and interpretation of data, and the preparation, review or approval of the manuscript.
Results
Of 3026 individuals enrolled in MOST, 30-month follow-up contact occurred in 2969 (see figure 1). Of the 57 with no 30-month contact, 33 had died and 24 could not be reached. Of the 2969 individuals with 30-month contact, 2713 completed both telephone interview and clinic visit, 215 completed only the telephone interview and a missed clinic visit telephone interview and 41 completed only the telephone interview. The reasons for not completing the 30-month clinic visit were: too busy (77 individuals); health problems (70); caregiving responsibilities (31); deceased (30); clinic too far (21); moved out of area (20); not satisfied with study (19); unable to contact (16); refused to give reason (8); personal problems (7) and other reasons in the remaining 14. Those not completing the 30-month clinic visit did not differ in age, gender, or alignment distribution in the dominant knee (32% neutral, 47% varus, 21% valgus) but had a higher BMI (32.0±6.9, SD vs 30.6±5.8) than those who completed this visit.
How the samples of 2958 knees for analyses of incident knee osteoarthritis (OA) and 1307 knees for analyses of knee osteoarthritis progression were derived. TKR, total knee replacement.
In the 2713 individuals who completed the 30-month clinic visit, 621 knees were excluded from analysis for a total knee replacement at baseline or at follow-up or for advanced osteoarthritis at baseline that could not progress further. In these 2713 individuals, 1614 right knees and 1692 left knees did not have osteoarthritis at baseline and were at risk of developing incident osteoarthritis. Of these 3306 knees, 348 were excluded for missing data (predominantly strength or laxity), resulting in a sample of 2958 knees for analyses of incident tibiofemoral osteoarthritis. Osteoarthritis was present at baseline in 797 right and 702 left knees; these knees were at risk of osteoarthritis progression. Of these 1499 knees, 192 were excluded for missing data, resulting in a sample of 1307 knees for analyses of osteoarthritis progression.
The 1752 participants who contributed 2958 knees for analyses of incident osteoarthritis had a mean age of 61.3 years (±7.8 SD), a mean BMI of 29.5 kg/m2 (±5.1) and included 1034 (59%) women. The baseline characteristics of these knees are summarised in table 1. The 950 participants who contributed the 1307 knees for analyses of osteoarthritis progression had a mean age of 63.6 years (±7.8), a mean BMI of 31.7 kg/m2 (±5.9) and included 592 (62%) women. The characteristics of these knees are summarised in table 2. In total, 4265 knees in 2287 individuals were analysed; 415 individuals contributed to both analysis samples.
Characteristics of knees without radiographic tibiofemoral osteoarthritis
Characteristics of knees with radiographic tibiofemoral osteoarthritis
As shown in table 3, the odds of developing incident tibiofemoral osteoarthritis were significantly elevated in knees with varus, but not in knees with valgus alignment at baseline, compared with neutral knees, in analyses adjusting for age, gender, BMI, knee injury, laxity and strength. The magnitude of the OR was comparable in the smaller strata of women and men considered separately. We secondarily analysed alignment as a continuous variable. Including only neutral and varus knees, greater severity of varus at baseline was associated with greater odds of incident osteoarthritis approaching significance (adjusted OR 1.07/1° varus; 95% CI 0.99 to 1.16). Among valgus and neutral knees, greater severity of valgus was not associated with greater odds of incident osteoarthritis (adjusted OR 0.98/1° valgus; 95% CI 0.85 to 1.14).
Varus and valgus alignment and incident radiographic tibiofemoral osteoarthritis
As shown in table 4, the odds of medial osteoarthritis progression were significantly elevated in knees that were varus at baseline. The medial osteoarthritis progression risk was significantly reduced in valgus knees. On the other hand, the odds of lateral osteoarthritis progression were significantly elevated in knees that were valgus at baseline and significantly reduced in varus knees (see table 4). The results were similar in men and women considered separately. The finding for valgus alignment and lateral progression in men may reflect that there were only 28 valgus knees with osteoarthritis in men. We secondarily analysed alignment as a continuous variable. Including only neutral and varus knees, greater severity of varus at baseline was significantly associated with greater odds of medial osteoarthritis progression (adjusted OR 1.29/1° varus; 95% CI 1.22 to 1.37). Among valgus and neutral knees, greater severity of valgus was significantly associated with greater odds of lateral osteoarthritis progression (adjusted OR 1.47/1° valgus; 95% CI 1.30 to 1.65).
Varus and valgus alignment and radiographic tibiofemoral osteoarthritis progression
Further adjustment for baseline K/L grade, baseline knee pain severity and concurrent change in BMI had minimal impact on these results.
Discussion
Varus but not valgus alignment increased the risk of incident radiographic tibiofemoral osteoarthritis. In the more vulnerable milieu of the knee with established osteoarthritis, varus and valgus alignment each increased the risk of osteoarthritis progression in the biomechanically stressed compartment and reduced the risk of progression in the unloaded compartment. A substantial proportion of knees were varus or valgus: 41% and 19%, respectively, in knees without tibiofemoral osteoarthritis; 58% and 18% in knees with osteoarthritis.
The pattern of results in the stressed and unloaded tibiofemoral compartments further supports that the mechanism of action of malalignment relates to its effect on load distribution. Varus alignment shifts the load-bearing axis medial to the knee centre, creating a moment arm that increases forces across the medial compartment and reduces lateral load; the lateral shift of the load-bearing axis due to valgus alignment increases forces across the lateral compartment and reduces medial load.
We found that varus but not valgus alignment increased the risk of incident osteoarthritis. Similarly, Brouwer et al5 observed that varus had a significant effect, whereas valgus had a borderline effect, in a study using a comparable definition of varus and valgus and defining, as we did, neutral knees as reference. The relationship between alignment and knee osteoarthritis development was also examined in a case–control study involving Framingham cohort members; in that study, the most varus (1–7° varus) was compared not with neutral knees but with the most valgus quartile (5–10° valgus) as referent,7 modelling a different question than we posed.
From a biomechanical perspective, a stronger finding for varus alignment is not surprising. Due to a stance phase knee adduction moment, greater load passes medially than laterally even in neutrally aligned, healthy knees.20 21 The adduction moment magnitude increases as varus alignment increases.22 Adduction moment magnitude predicted knee osteoarthritis progression23; an adduction moment increase may lie in the causal pathway between varus alignment and knee osteoarthritis progression. Varus alignment further increases the total load passing medially.24 Although valgus alignment is associated with an increase in lateral compartment peak pressures,25 the medial compartment often continues to bear more load until more severe valgus is present.26 27
Alternatively, the inability to detect an association between valgus alignment and incident osteoarthritis may reflect a lower sensitivity of the measure of incident osteoarthritis compared with the measure of osteoarthritis progression. In the vast majority of knees, osteophyte development precedes joint space narrowing. The definition of incident osteoarthritis hinges upon these osteophytes. Because osteophyte formation is neither specific to the involved nor to the spared compartment, at early stages of osteoarthritis, radiographs cannot reveal whether a knee has medial or lateral osteoarthritis. It is possible that the greater overall frequency of medial compared with lateral osteoarthritis dilutes the detected valgus effect upon a non-compartment-specific measure of incident osteoarthritis. In contrast, the measure of progression allows specific examination of the compartment stressed by the alignment (medial for varus, lateral for valgus). As such, in the analysis of the impact of alignment, the measure of incident osteoarthritis, which is not compartment specific, is generally inferior to the measure of progression.
In the analyses of knees with osteoarthritis at baseline, varus and valgus alignment were associated with medial and lateral osteoarthritis progression, respectively, in keeping with previous studies.2,–,6 Brouwer et al5 also found significant results for varus only; the association between valgus and lateral progression was significant only for those who were obese.5 As the authors note, the inability to detect the valgus effect may relate to the use of K/L worsening to define progression and the relatively small numbers with progression.
The source of malalignment predating knee osteoarthritis may be genetic, developmental, or traumatic. Before the study by Brouwer et al,5 evidence that malalignment may contribute to osteoarthritis development came from animal models and human fracture studies. Our findings and those of Brouwer et al5 support that, in knees without radiographic tibiofemoral osteoarthritis, varus alignment increases the risk of osteoarthritis development. The prevalence of varus and valgus alignment in knees without radiographic osteoarthritis—that is, before any loss of bone and cartilage height that could contribute to malalignment—supports that not all malalignment is a consequence of disease. Whatever the original cause of the alignment, varus and valgus alignment each increase the risk of subsequent knee osteoarthritis progression. It seems likely that the worsening of tibiofemoral osteoarthritis in turn increases malalignment, at least in some knees. With only one time point of alignment measurement, we could not explore this. The existence of a vicious cycle does not lessen the impact of our findings; strategies that interrupt a vicious cycle may be a potent means of delaying progression. These results support the further development and testing of non-invasive modalities to improve tibiofemoral load distribution in varus-aligned and valgus-aligned knees.
It is important to acknowledge that MOST participants without knee osteoarthritis were at greater risk of developing it. Those at greater risk of developing knee osteoarthritis are of particular public health importance,9 and it is crucial to understand the relationship of alignment to incident osteoarthritis in them. Varus alignment was associated with incident osteoarthritis by the established definition. We lacked sufficient power to examine incidence within K/L 0 and 1 strata separately. With or without these analyses, it cannot be concluded from any radiographic study that any factor initiates knee osteoarthritis, given the insensitivity of x-rays to early osteoarthritis pathology and the inability to identify the point of osteoarthritis onset. Although there is no consensus as yet about what knee MRI feature(s) constitute osteoarthritis, future studies should explore the relationship between alignment and osteoarthritis development in knees without MRI-based measures of osteoarthritis pathology. We were not able to assess the change in alignment between baseline and follow-up. Those who did not complete the 30-month visit had a higher BMI; it is uncertain what impact this may have had on the results.
In conclusion, varus but not valgus alignment increased the risk of incident radiographic tibiofemoral osteoarthritis. In knees with established osteoarthritis, varus and valgus alignment each increased the risk of osteoarthritis progression in the biomechanically stressed compartment.
References
Footnotes
-
Funding This study was funded by NIH NICHD RO1 HD43500; NIA U01 AG18820, AG18832, AG18947, AG19069.
-
Competing interests None.
-
Ethics approval The study protocol was approved by the institutional review boards at each participating site.
-
Provenance and peer review Not commissioned; externally peer reviewed.