Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis Monica R. Maly a,n, Stacey M. Acker b, Saara Totterman c, José Tamez-Peña c,d, Paul W. Stratford a, Jack P. Callaghan b, Jonathan D. Adachi e, Karen A. Beattie e a
School of Rehabilitation Sciences, McMaster University, Hamilton, Canada Department of Kinesiology, University of Waterloo, Waterloo, Canada c Qmetrics Technologies, Rochester, NY, USA d Escuela de Medicina, Tecnológico de Monterrey, Monterrey, NL, México e Department of Medicine, McMaster University, Hamilton , Canada b
art ic l e i nf o
a b s t r a c t
Article history: Accepted 27 April 2015
The objective was to determine the extent to which the external peak knee adduction moment (KAM) and cumulative knee adductor load explained variation in medial cartilage morphology of the tibia and femur in knee osteoarthritis (OA). Sixty-two adults with clinical knee OA participated (61.576.2 years). To determine KAM, inverse dynamics was applied to motion and force data of walking. Cumulative knee adductor load reflected KAM impulse and loading frequency. Loading frequency was captured from an accelerometer. Magnetic resonance imaging scans were acquired with a coronal fat-saturated sequence using a 1.0 T peripheral scanner. Scans were segmented for medial cartilage volume, surface area of the bone–cartilage interface, and thickness. Forward linear regressions assessed the relationship of loading variables with cartilage morphology unadjusted, then adjusted for covariates. In the medial tibia, age and peak KAM explained 20.5% of variance in mean cartilage thickness (p o0.001). Peak KAM alone explained 12.3% of the 5th percentile of medial tibial cartilage thickness (i.e., thinnest cartilage region) (p ¼0.003). In the medial femur, sex, BMI, age, and peak KAM explained 44% of variance in mean cartilage thickness, with peak KAM contributing 7.9% (po0.001). 20.7% of variance in the 5th percentile of medial femoral cartilage thickness was explained by BMI and peak KAM (p¼ 0.001). In these models, older age, female sex, greater BMI, and greater peak KAM related with thinner cartilage. Models of KAM impulse produced similar results. In knee OA, KAM peak and impulse, but not loading frequency, were associated with cartilage thickness of the medial tibia and femur. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Arthritis Cartilage Articular Knee joint Locomotion
1. Introduction Joint loading is implicated in knee osteoarthritis (OA) progression (Andriacchi et al., 2009). Over six years, an increase by 25% of the peak knee adduction moment (KAM) corresponded with a 6.46 times greater risk for progressive joint space narrowing in knee OA (Miyazaki et al., 2002). Similar findings have been produced with magnetic resonance imaging (MRI). In 180 people with medial knee OA, more severe cartilage defects were associated with higher peak KAM (Creaby et al., 2010). Five-year changes in medial tibial cartilage thickness were predicted by baseline peak KAM and peak knee flexion moment in 16 people with knee OA (Chehab et al., 2014). However, the peak KAM does n Correspondence to: McMaster University, Room 435 IAHS, 1400 Main St. West, Hamilton, ON, Canada L8S 1C7. Tel.: þ 1 905 525 9140x27823; fax: þ1 905 524 0069. E-mail address: [email protected] (M.R. Maly).
not always relate with cartilage damage. Peak KAM was unrelated to progression of cartilage defects over 12 months in medial knee OA (Bennell et al., 2011). It may be unreasonable to expect that a laboratory measure captured during a single instant would share a relationship with cartilage features that reflect long-term loading. The magnitude, duration and repetition of weight-bearing activities dictate the responses of cartilage to mechanical loads (Pearle et al., 2005; Maly et al., 2012). Compared to peak KAM, the combination of KAM impulse and loading frequency encountered in daily activity may better reflect cartilage morphology. Accounting for both KAM impulse and loading frequency together explained variance in pain among adults with knee OA (Robbins et al., 2011a, 2011b) and was superior to the peak KAM in distinguishing between adults with and without knee OA (Maly et al., 2012). KAM impulse reflects the total duration and magnitude of medial knee loading during one stride. KAM impulse was associated with pain intensity in knee OA (Thorp et al., 2007; Robbins et al., 2011a, 2011b). In radiographic knee OA, this moment-time
http://dx.doi.org/10.1016/j.jbiomech.2015.04.039 0021-9290/& 2015 Elsevier Ltd. All rights reserved.
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
integral distinguished between moderate and mild radiographic disease; whereas the peak KAM was not different between groups (Thorp et al., 2006). A higher KAM impulse, but not peak KAM, was associated with greater loss of medial tibial cartilage volume over 12 months in knee OA (Bennell et al., 2011). Loading frequency is also important in cartilage degradation. Epidemiological studies of sport and occupation show a dose– response relationship between loading frequency and knee OA incidence (Coggon et al., 2000; Vignon et al., 2006; Stehling et al., 2010). A 10-year follow-up of 16,961 patients demonstrated that high loading frequency, noted by distance of a walk or run, had an odds ratio of 2.4 (95%CI 1.5–3.9) for symptomatic knee OA in men (Cheng et al., 2000). Greater steps/day increased cartilage loss over 2.7 years in those with low cartilage volume at baseline; but interestingly protected against cartilage loss in those with greater cartilage volume at baseline (p ¼0.046) (Dore et al., 2013). High loading frequency (4 10,000 steps/day) was associated with a greater severity of cartilage defects in this sample (Dore et al., 2013). The purpose of this study was to determine the extent to which the peak KAM, and cumulative knee adductor load (KAM impulse, loading frequency) explained variation in medial tibial and femoral cartilage morphology in knee OA. We hypothesized that peak KAM would relate inversely with the 5th percentile of cartilage thickness; while elements of cumulative load would show stronger relationships with morphology. Specifically, we hypothesized that KAM impulse would relate inversely with cartilage volume and thickness, and positively with surface area of the bone–cartilage interface; while loading frequency would relate inversely with cartilage volume. 2. Methods This cross-sectional study was approved by the Institutional Human Research Ethics Board.
Table 1 Nomenclature for measures of cartilage morphology are consistent with published norms (Eckstein et al., 2006a). The medial tibia and medial femur were represented by the prefixes MT and MF respectively. VC refers to volume of cartilage. tAB refers to total area of subchondral bone. cAB refers to area of subchondral bone covered in cartilage. Finally, ThCtAB refers to cartilage thickness over total subchondral bone area. All measures represent mean values from the medial compartment of the knee.
Cartilage volume (mm3) Surface area of bone–cartilage interface (mm2) Cartilage thickness (mm)
Tibial
Femoral
MT.VCtAB MT.cAB MT.ThCtAB
MF.VCtAB MF.cAB MF.ThCtAB
at the femoral trochlea in images of OA knees obtained with a 3.0 T MRI (TamezPena et al., 2012) and has evaluated cartilage repair (Shive et al., 2014) and bone morphology (Hunter et al., 2014). The accuracy of segmentation from scans acquired with a 1.5 T magnet produced data within 4–6% of values from a 3.0 T (Schneider et al., 2012). No statistical differences in cartilage measurements existed between 1.5 T and 1.0 T magnets (Inglis et al., 2007). Briefly, eight atlases were used to segment each scan. Each atlas was morphed to match anatomy using a spline deformation with a mutual information similarity metric. After spline morphing, the edges of the segmented image were deformed using a free-form registration algorithm. The morphed cartilage segmentation was then corrected by outlier detection: voxels whose signal was three standard deviations higher than average were classified as outliers. Segmentations from the eight atlases were fused using a fuzzy voting algorithm, and statistically relaxed to create the final segmentation (Tamez-Pena et al., 2012). All segmentations were reviewed by an experienced radiologist (ST) for quality. The following were calculated from the medial tibia and femur: volume (VCtAB; mm3), surface area of the bone–cartilage interface (cAB; mm2), mean thickness (ThCtAB; mm) and mean of the 5th percentile thickness (mm). These acronyms (Table 1) are consistent with established nomenclature (Eckstein et al., 2006a, 2006b). Coronal weight-bearing knee radiographs were obtained in a standardized fixed-flexion position using a SynaflexerTM (Kothari et al., 2004). This frame places the feet in 5° of external rotation and 20° of knee flexion. Each digital radiograph was evaluated by one experienced radiologist (ST) to yield Kellgren and Lawrence (K–L) scores (Kellgren and Lawrence, 1957) and frontal plane alignment (anatomical axis). 2.3. Biomechanical Assessment
2.1. Participants Community-dwelling adults between 40 and 70 years who met the American College of Rheumatology (ACR) clinical criteria for knee OA were recruited from rheumatologist's and orthopaedic surgeon's offices. These guidelines include having knee pain on most days of the month and at least three of the following: Z50 years of age; stiffness o30 min; crepitus; bony tenderness; bony enlargement; no palpable warmth (Altman et al., 1986). Exclusion criteria included other forms of arthritis; non-arthritic disease; intra-articular therapies; or previous knee surgeries (e.g. osteotomy, replacement, partial/complete menisectomy, ligament reconstruction). Potential participants were excluded if they required an adaptive walking aid; sustained leg trauma within the past 3 months; had ipsilateral hip or ankle conditions; or had contraindications to MRI. In those with bilateral knee OA, the knee with more severe symptoms was designated as the study knee. All participants provided written, informed consent. 2.2. Cartilage morphometry Each participant underwent an MRI scan of the study knee using a 1.0 T peripheral MRI scanner (GE Healthcare, USA). Participants were seated with the knee fully extended and centered in the iso-center of the 180 mm removable quadrature volume transmit-receive coil. Padding around the knee, thigh, and leg limited movement. Sagittal gradient-echo and axial fast spin-echo localizer scans were performed (2–3 min). A coronal fat-saturated spoiled gradient recalled acquisition in the steady-state (SPGR) was acquired: TR 60 ms; TE 12.4 ms (or minimum); flip angle 40°; bandwidth 30 kHz; matrix 512 256 (frequency phase); 1 excitation; field of view 150 mm; slice thickness 1.5 mm; 56–64 partitions depending on patient size. Scan time was 15–16 min. MRI fat-saturated sequences (i.e., SPGR) from similar lower level magnets quantify cartilage morphology with accuracy of 13% to 3% and precision around 4.0% (Eckstein et al., 1998; Stammberger et al., 1999; Burgkart et al. 2001; Eckstein and Glaser, 2004). Medial tibial and femoral cartilage morphology was segmented from these images using a highly automated, atlas-based method (Qmetrics, Rochester, NY, USA) (Tamez-Pena et al., 2012). This method yielded test–retest precision of cartilage thickness values between 0.014 mm (0.6%) at the femur and 0.038 mm (1.6%)
Within one week of the MRI, a laboratory visit was conducted. To calculate the peak and impulse of the external KAM, gait analyses were conducted during barefoot walking at a self-selected speed. Kinetics were collected using a synchronized floor-mounted force plate (OR6-7, Advanced Mechanical Technology Inc., Watertown, MA, USA) sampled at 1000 Hz. Concurrently, kinematics were collected using three Optotrak Certus banks (nine cameras) (Northern Digital Inc., Waterloo, ON, Canada) sampled at 100 Hz. Infrared emitting diodes, arranged on rigid bodies in clusters of three, were secured to the lateral sacrum, thigh, shank and foot. The pelvis was digitized using six landmarks (bilateral anterior and posterior superior iliac spines, greater trochanters). Virtual markers were also created for the greater trochanter, medial and lateral femoral and tibial condyles, tibial tuberosity, fibular head, medial and lateral malleoli, calcaneus, and first, second and fifth metatarsal heads. A static reference trial was recorded to determine neutral pelvic and lowerlimb joint angles. Participants ambulated until five trials were captured where the participant struck the force plate cleanly with the study leg. Using commercial software (Visual 3D, C-Motion, Inc., Germantown, MD, USA), marker data were dual-passed through a low-pass Butterworth filter with a 6 Hz cut-off frequency (Robertson and Dowling 2003). Inverse dynamics was used to calculate external knee moments (Winter 1990) in a three-dimensional floating axis coordinate system where flexion/extension occurred about the lateral-medial axis of the thigh, internal/external rotation occurred about the distal-proximal axis of the shank, and adduction/abduction occurred about a floating axis (intermediate axis perpendicular to both lateral-medial and distal-proximal axes) (Cole et al. 1993; Wu and Cavanagh 1995; Schache and Baker 2007). The rotation sequence was consistent with a Cardan XYZ rotation (Wu and Cavanagh 1995). The KAM peak and impulse are highly reliable, stable measures in knee OA (Birmingham et al. 2007; Robbins et al. 2009). Peak KAM was reported in Nm/kg (Winter 1990). Mean KAM impulse of five trials was calculated using the trapezoid rule, which integrated only positive values from stance (Matlab 7.0.1, Natick, MA, USA). Our interest focused on the medial knee; thus negative (abduction) values were not subtracted from the adduction impulse because these values do not subtract from the loading experienced by the medial knee. To reflect the theoretical underpinnings of cumulative load, which represents the total load exposure experienced by tissues for a given time period, KAM impulse values were in non-normalized units of Nm●s. Non-normalization ensures that the impulse
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ reflects the absolute cumulative knee load during stride (Browning and Kram 2007). Loading frequency was the daily average number of steps on the study leg using a tri-axial accelerometer (GTX3 þ, ActiGraph, Pensacola, Florida, USA) (Kochersberger et al. 1996; Gardiner et al. 2011). The sampling rate was 30 Hz. Participants wore the accelerometer for waking hours, except for activities involving water, for seven consecutive days. The accelerometer was secured on a waist-belt, directly above the thigh midline of the study leg. Participants completed an activity logbook to record periods of non-wear. Any day where the accelerometer was not worn for least 10 hours was excluded (Song et al. 2010). Accelerometry in elderly samples required five days to estimate daily physical activity (Trost et al. 2005). Step counts from the first five full days were averaged. 2.4. Data analyses A series of analyses were completed to build linear models that expressed the relationship of mechanics with medial cartilage morphology. First, forward linear regressions were computed where measures of medial cartilage morphology were the dependent variables: volume, surface area of the bone-cartilage interface, mean thickness, and 5 th percentile thickness of the tibia and femur. Two models of independent variables were composed. In Model 1, the relationship of peak KAM with tibial and femoral cartilage morphology was explored. In Model 2, relationships of cartilage morphology with KAM impulse and steps/day were explored together, representing cumulative knee adductor load (Robbins et al. 2011a; 2011b; Maly et al. 2012). These models were explored unadjusted. Next, adjusted forward regressions were computed using covariates of age, sex, and BMI, provided these were significant (i.e., p o0.05). Older age, female sex and greater BMI were associated with the development and progression of cartilage defects over two years (Ding et al. 2006) and greater cartilage loss was observed in women than men, particularly with aging (Ding et al. 2007). Finally, if the independent variables of interest (e.g., peak KAM, KAM impulse, loading frequency) were not entered in the adjusted model, we conducted a follow-up analysis that included statistically significant covariates and independent variables of interest to clearly show that the independent variables of interest did not contribute to the model. The probability of F statistic was set at 0.05 to enter. Statistical significance was set at P o 0.05. Histograms of the residuals and plots of the residuals against the predicted values of the dependent variables were utilized to examine the appropriateness of the regression analysis. Tolerance (Z 0.908) and variance inflation factors ( r 1.1) demonstrated that multicollinearity was not present. Analyses were completed using SPSS 21.0 (SPSS Inc., Chicago, USA).
3. Results Sixty-two adults (11 men) participated (Table 2). This group included 27, 18 and 17 participants with K-L scores of 2, 3 and 4, respectively. All had radiographic medial OA. Anatomical alignment was 3.4 72.6° varus (range¼-5.9 to 11.3°) with 19 of 62 participants demonstrating valgus (alignment o0°). Table 2 Descriptive statistics including demographics, gait mechanics and cartilage morphology (n¼ 62). Variable
Mean (Standard Deviation) Minimum– maximum
Age (y) Body Mass (kg) Height (m) Body Mass Index (kg/m2) Peak KAM (Nm/kg) KAM Impulse (Nm●s) Load Frequency (steps/day)a MT.VCtAB (mm3) MT.cAB (mm2) MT.ThCtAB (mm) Tibial 5th% Thickness (mm) MF.VCtAB (mm3) MF.cAB (mm2) MF.ThCtAB (mm) Femoral 5th% Thickness (mm)
61.5 (6.2) 74.1 (16.0) 1.61 (0.10) 28.2 (5.7) 0.35 (0.17) 9.14 (6.29) 3898 (1951) 1652 (384) 918 (148) 1.8 (0.2) 0.8 (0.1) 1740 (398) 877 (114) 1.8 (0.2) 0.8 (0.1)
41–70 49.4–117.0 1.07–1.85 19.7–41.8 0.01–0.72 0.87–26.94 644–10968 934–2929 591–1463 1.3–2.3 0.5–1.0 1201–3225 657–1188 1.5–2.5 0.6–0.9
KAM is the external knee adduction moment. a
Steps/day reflect loading frequency for the study leg only.
3
3.1. Tibial cartilage morphology Table 3 summarizes the models for the medial tibia. Loading frequency shared a positive relationship with MT.VCtAB; adjusting for sex and age eliminated this relationship. Both KAM impulse and loading frequency were positively related to MT.cAB. However adjusting for demographics eliminated these from the final model. Thus, male sex was associated with larger cartilage volume and surface area; and young age related to greater medial tibial cartilage volume. Cartilage thickness was inversely related with KAM. Peak KAM explained 6.6% of the variance in MT.ThCtAB after adjusting for age (p o0.001). The KAM impulse explained 5.1% of the variance in MT.ThCtAB after adjusting for age and sex (p o0.001). Beta coefficients for the adjusted peak KAM, in Model 1, and KAM impulse, in Model 2, were not appreciably different from the unadjusted values. Peak KAM alone explained 12.3% of the variance in the mean 5th percentile of cartilage thickness in the medial tibia (p ¼0.003), such that a greater peak KAM related with thinner cartilage. KAM impulse explained 15.6% of variance in the mean 5th percentile of cartilage thickness in the medial tibia (p o0.005). Loading frequency did not contribute to variance in tibial cartilage morphology in Model 2 after adjusting for age, sex and BMI. 3.2. Femoral cartilage morphology Table 4 summarizes the models for the medial femur. Male sex was associated with larger MF.VCtAB and a larger MF.cAB; and young age related to greater MF.VCtAB. Peak KAM explained 7.9% of the variance in MF.ThCtAB, where a greater peak KAM related with thinner cartilage, after adjusting for sex, BMI and age (p o0.001). Similarly, 8.4% of the variance in MF.ThCtAB was explained by the KAM impulse after adjusting for sex, BMI and age (p o0.001). BMI and the peak KAM explained 20.7% of the variance in the mean 5th percentile of cartilage thickness in the medial femur, with peak KAM contributing 12.5% (p o0.001). Both BMI and peak KAM inversely related with mean 5th percentile of cartilage thickness. KAM impulse explained 10.5% of variance in the mean 5th percentile of medial femoral cartilage thickness after adjusting for BMI (p o0.001). Beta coefficients for the unadjusted and adjusted peak KAM and KAM impulse were similar. Loading frequency did not contribute to variance in any of the femoral cartilage morphology measures in Model 2.
4. Discussion In clinical, radiographic knee OA, the peak KAM and KAM impulse explained variance in cartilage thickness of the medial tibia and femur. Greater medial knee loading was associated with thinner cartilage. No relationship existed between loading frequency and cartilage morphology. These findings highlight the importance of biomechanics in cartilage degradation by demonstrating that medial knee loads are associated with cartilage morphometry in vivo. A novel contribution is exploring loading frequency in the relationships of biomechanics with cartilage morphology in knee OA. Loading frequency was not related to cartilage morphology after accounting for demographics. Thus, the cumulative knee adductor load approach was not useful. A cross-sectional snapshot of loading frequency may not reflect an individual’s lifespan of physical activity. Previous work using a variety of methods to characterize loading frequency and knee health have yielded conflicting results (Urquhart et al., 2008). Mediators of this relationship include sex, age, BMI, body composition, strength, injury, and gait biomechanics (Urquhart et al., 2008). People who have
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
Table 3 Forward linear regression models of medial tibial cartilage morphology. Model 1 includes the peak knee adduction moment (KAM) as a potential predictor. Model 2 includes the KAM impulse and loading frequency as potential predictors. The models are presented unadjusted and adjusted for age, sex and BMI. The final model is presented in bold. Dependent variable
Model 1 Predictors
MT.VCtAB (mm3)
MT.cAB (mm2)
MT.ThCtAB (mm)
Unadjusted Peak KAM
Adjusted R2
Unstandardized β coefficient
P
0.01
232.936
0.429
Adjusted Sex þ Age Sex Age Peak KAM
0.444 0.502 0.509
643.957 15.764 272.254
0.001 0.001 0.007 0.186
Unadjusted Peak KAM
0.02
29.862
0.793
Adjusted Sex Peak KAM
0.499 0.488
280.746 4.540
0.001 0.955
Unadjusted Peak KAM
0.068
0.351
0.023
Adjusted Age þ Peak KAM Age Peak KAM
5th Percentile tibial thickness (mm)
Model 2
Unadjusted Peak KAM
Adjusted Peak KAM
0.001 0.139 0.205
0.123
0.123
0.011 0.325
0.205
0.205
0.003 0.019
0.003
0.003
greater step counts may possess other features, such as greater fitness, that may protect cartilage health but were not captured in this study. Baseline knee health also appears important. Among 405 adults between 51 and 81 years of age, high frequency loading protected against cartilage loss over 2.7 years in those with large cartilage volumes at baseline; but facilitated cartilage loss in those with smaller cartilage volume at baseline (Dore et al., 2013). Mechanical loads of magnitudes beyond the physiological capacity of cartilage facilitate degeneration of this tissue (Felson et al., 2000; Miyazaki et al., 2002; Urquhart et al., 2008; Bennell et al., 2011). Our findings identify that large KAMs during walking were associated with the presence of thinned tibiofemoral cartilage. While sex, age and BMI were associated with cartilage morphology in some models, the consistency of the beta coefficients for KAM peak and impulse in the unadjusted and adjusted models suggests that these covariates did not alter the association between medial loading and cartilage thickness of the medial tibia and femur. Previous work has produced conflicting results. While a high KAM peak and impulse related to meniscal extrusion (p o0.001) and lower medial meniscus height (p¼ 0.010), the
Predictors
Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Sex þ Age Sex Age KAM impulse Loading frequency Unadjusted KAM impulse þLoading frequency KAM impulse Loading frequency Adjusted Sex KAM impulse Loading frequency Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Age þ Sex þ KAM impulse
Adjusted R2
Unstandardized β coefficient
P
0.018 0.013 0.104
11.436 0.065
0.152 0.012
0.444 0.502 0.493 0.499
643.957 15.764 0.696 0.025
0.001 0.001 0.003 0.911 0.210
0.004 0.085 0.154
7.826 0.022
0.010 0.021
0.499 0.497 0.513
280.746 2.794 0.012
0.001 0.242 0.102
0.080 0.006 0.055
0.005 0.001
0.256 0.055 0.001
Age Sex
0.150 0.202
0.011 0.175
0.006 0.009
KAM impulse Loading frequency
0.253 0.245
0.008 0.001
0.035 0.510
0.156
0.006
0.005
0.129
0.001
0.211
0.156 0.129
0.006 0.001
0.005 0.211
Unadjusted KAM impulse Loading frequency Adjusted KAM impulse Loading frequency
thickness of cartilage in the tibia and femur were unrelated to the KAM peak and impulse in 45 women with clinical knee OA (Vanwanseele et al., 2010). Another study conducted cross-sectional analyses of 180 people with radiographic knee OA, which showed that both the peak KAM and KAM impulse were positively associated with medial knee cartilage defect scores (Creaby et al., 2010). Sample size and differences in measurement (e.g. cartilage thickness from segmentation versus categorization of cartilage defects) may be responsible for this discrepancy. The 5th percentile of cartilage thickness quantified the average thickness of thinnest cartilage tissue. This measure likely equates with previous categorizations of cartilage defects (Creaby et al., 2010; Bennell et al., 2011). Several methodological differences exist between Creaby’s work (Creaby et al., 2010) and the present study. Creaby et al. (2010) recruited a narrower range of radiographic severities of knee OA (mild–moderate), analyzed shod gait, and utilized manual segmentation from MRI scans from two different 1.5T whole-body scanners. The current study recruited people with clinical knee OA (K–L 2–4), analyzed barefoot gait and used highlyautomated segmentation of scans from one 1.0 T peripheral
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Table 4 Forward linear regression models of medial femoral cartilage morphology. Model 1 includes the peak knee adduction moment (KAM) as a potential predictor. Model 2 includes the KAM impulse and loading frequency as potential predictors. The models are presented unadjusted and adjusted for age, sex and BMI. The final model is presented in bold. Dependent variable
Model 1 Predictors
MF.VCtAB (mm3)
MF.cAB (mm2)
MF.ThCtAB (mm)
5th Percentile femoral thickness (mm)
Unadjusted Peak KAM
Model 2 Adjusted R2
Unstandardized β coefficient
P
0.005
329.384
0.257
Adjusted Sex þ Age Sex Age Peak KAM
0.415 0.454 0.473
628.784 13.360 370.354
0.001 0.001 0.027 0.085
Unadjusted Peak KAM
0.017
4.715
0.957
Adjusted Sex Peak KAM
0.439 0.433
205.643 20.409
0.001 0.756
Unadjusted Peak KAM
0.077
0.322
0.017
Adjusted Sex þ BMI þ Age þ Peak KAM Sex BMI Age Peak KAM
0.170 0.269 0.362 0.441
0.191 0.011 0.008 0.310
0.001 0.001 0.006 0.004
Unadjusted Peak KAM
0.124
0.126
0.003
Adjusted BMI þ Peak KAM BMI Peak KAM
0.001
0.001 0.082 0.207
0.003 0.125
scanner. The consistency in findings despite methodological differences provides confidence that the relationships between walking mechanics and cartilage defects exist. The peak KAM, which lasts a fraction of one second during stride, and the KAM impulse, which reflects both the duration and magnitude of medial loads throughout stride, produced similar results. (Vanwanseele et al., 2010) noted a similar pattern in 45 women with clinical knee OA. The distinction between KAM peak and impulse may not be important to cartilage morphology. Instead, a general picture of load magnitude appeared sufficient to relate with cartilage thickness. The peak KAM was less sensitive than the KAM impulse when distinguishing between radiographic severities of knee OA (Thorp et al., 2006; Robbins et al., 2011a; 2011b; Kean et al., 2012). Cross-sectionally, both the peak KAM and KAM impulse were positively associated with severity of cartilage defects (Creaby et al., 2010). However, follow-up on the
0.011 0.002
Predictors
Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Sex þ Age Sex Age KAM impulse Loading frequency Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Sex KAM impulse Loading frequency
Adjusted R2
P
0.201 0.018 0.022
1.330 0.047
0.870 0.075
0.413 0.452 0.477 0.469
630.332 13.646 11.887 0.008
0.001 0.001 0.031 0.069 0.699
0.064 0.050 0.074
4.864 0.012
0.044 0.124
0.440 0.442 0.436
207.039 0.971 0.004
0.001 0.621 0.514
Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Sex þ BMI þ Age þ KAM impulse
0.053 0.095
Sex BMI Age KAM impulse Loading frequency
0.170 0.263 0.360 0.444 0.435
0.024 0.008 0.001
0.045 0.062 0.001
Unadjusted KAM impulse þ Loading frequency KAM impulse 0.164 Loading frequency 0.149 Adjusted BMI þ KAM impulse BMI KAM impulse Loading frequency
Unstandardized β coefficient
0.107 0.212 0.183
0.237 0.008 0.008 0.010 0.001
0.001 0.035 0.012 0.004 0.701
0.004 0.004 0.001
0.001 0.880 0.001
0.003 0.003 0.001
0.051 0.006 0.791
same sample showed that over one year, only KAM impulse was related to cartilage volume loss (Bennell et al., 2011). It is possible that changes in cartilage volume over one year were too small to observe this association. Surface area of the bone–cartilage interface was unrelated to biomechanical measures, in contrast to previous work. Creaby et al. (2010) showed that loading may have a role in widening and flattening the tibial plateau. Large mechanical loads attributed to training (triathletes) and obesity related to an expansion of tibial surface area (Ding et al., 2005). Subchondral bone area was directly related to the peak KAM (Jackson et al., 2004; Dempsey et al., 2013) and related to the ratio of medial to lateral tibial subchondral bone area (Vanwanseele et al., 2010). The discrepancy between findings in the current study and previous work may reflect definitions characterizing surface area. We quantified areas of bone that were covered in cartilage; whereas previous work
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
identified the subchondral bone surface area, regardless of the presence of cartilage (Jackson et al., 2004; Vanwanseele et al., 2010). Finally, cartilage volumes of the medial tibia and femur were unrelated to the biomechanical measures. It is likely that regionspecific analyses, such as those proposed by Eckstein et al. (2006a, 2006b) are necessary. Regional variations in knee cartilage thickness and volume are associated with loading patterns during gait (Koo and Andriacchi, 2007). High loads at the knee occur at heelstrike where the knee is near full extension (Koo et al., 2011) when the thickest regions of femoral and tibial cartilage are aligned (Andriacchi et al., 2004; Eckstein et al., 2006a; 2006b). The cross-sectional nature of this study limits the conclusions to associations only. While all participants had radiographic evidence of medial disease, we did not exclude those with concurrent lateral disease. Also, the anatomical axis angles for this sample suggest that participants were either varus, or demonstrated slight valgus. Together these data suggest that the participants likely loaded the medial knee to a greater extent than the lateral. Finally, eight dependent variables were tested. Multiplicity may have increased the likelihood of identifying statistically significant relationships. 4.1. Conflicts of interest statement Saara Totterman and José Tamez-Peña are employed by QMetrics Technologies.
Acknowledgements Funded by Canadian Institutes of Health Research (CIHR) Operating Grant (#102643 MRM). Salary support was provided by a CIHR New Investigator Award (MRM), Tier I Canada Research Chair (JPC) and Arthritis Society/Canadian Arthritis Network Scholar Award (KAB). Equipment required to conduct this research was purchased with the support of the Canadian Foundation for Innovation Leaders Opportunity Fund and the Ministry of Research and Innovation – Ontario Research Fund (MRM).
References Altman, R., Asch, E., Bloch, D., Bole, G., Borenstein, D., Brandt, K., Christy, W., Cooke, T., Greenwald, R., Hochberg, M., Howell, D., Kaplan, D., Koopman, W., Longley, S., Mankin, H., McShane, D., Medsger, T., Meenan, R., Mikkelsen, W., Moskowitz, R., Murphy, W., Rothschild, B., Segal, M., Sokoloff, L., Wolfe, F., 1986. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Arthritis Rheum. 29 (8), 1039–1049. Andriacchi, T., Koo, S., Scanlan, S., 2009. Gait mechanics influence healthy cartilage morphology and osteoarthritis of the knee. J. Bone Jt. Surg. Am. 91, 95–101. Andriacchi, T., Mundermann, A., Smith, R., Alexander, E., Dyrby, C., Koo, S., 2004. A framework for the in vivo pathomechanics of osteoarthritis at the knee. Ann. Biomed. Eng. 32, 447–457. Bennell, K., Bowles, K., Wang, Y., Cicuttini, F., Davies-Tuck, M., Hinman, R., 2011. Higher dynamic medial knee load predicts greater cartilage loss over 12 months in medial knee osteoarthritis. Ann. Rheum. Dis. 70, 1770–1774. Birmingham, T., Hunt, M., Jones, I., Jenkyn, T., Giffin, J., 2007. Test–retest reliability of the peak knee adduction moment during walking in patients with medial compartment knee osteoarthritis. Arthritis Rheum. 57, 1012–1017. Browning, R., Kram, R., 2007. Effect of obesity on the biomechanics of walking at different speeds. Med. Sci. Sports Exerc. 39, 1632–1641. Burgkart, R., Glaser, C., Hyhlik-Durr, A., Englmeier, K., Reiser, M., Eckstein, F., 2001. Magnetic resonance imaging-based assessment of cartilage loss in severe osteoarthritis: Accuracy, precision, and diagnostic value. Arthritis Rheum. 44, 2072–2077. Chehab, E., Favre, J., Erhart-Hledik, J., Andriacchi, T., 2014. Baseline knee adduction and flexion moments during walking are both associated with 5 year changes in patients with medial knee osteoarthritis. Osteoarthr. Cartil. 22, 1833–1839. Cheng, Y., Macera, C., Davis, D., Ainsworth, B., Troped, P., Blair, S., 2000. Physical activity and self-reported, physician-diagnosed osteoarthritis: is physical activity a risk factor? J. Clin. Epidemiol. 53, 315–322.
Coggon, D., Croft, P., Kellingray, S., Barrett, D., McLaren, M., Cooper, C., 2000. Occupational physical activities and osteoarthritis of the knee. Arthritis Rheum. 43, 1443–1449. Cole, G., Nigg, B., Ronsky, J., Yeadon, M., 1993. Application of the joint coordinate system to three-dimensional joint attitude and movement representation: a standardization protocol. J. Biomech. Eng. 115, 344–349. Creaby, M., Wang, Y., Bennell, K., Hinman, R., Metcalf, B., Bowles, K., Cicuttini, F., 2010. Dynamic knee loading is related to cartilage defects and tibial plateau bone area in medial knee osteoarthritis. Osteoarthr. Cartil. 18, 1380–1385. Dempsey, A., Wang, Y., Thorlund, J., Mills, T., Bennell, K., Metcalf, B., Hanna, F., Cicuttini, F., Lloyd, D., 2013. The relationship between patellofemoral and tibiofemoral morphology and gait biomechanics following arthroscopic medial meniscectomy. Knee Surg. Sports Traumatol. Arthrosc. 21, 1097–1103. Ding, C., Cicuttini, F., Blizzard, L., Scott, F., Jones, G., 2007. A longitudinal study of the effect of sex and age on rate of change in knee cartilage volume in adults. Rheumatology 46, 273–279. Ding, C., Cicuttini, F., Scott, F., Boon, C., Jones, G., 2005. Association of prevalent and incident knee cartilage defects with loss of tibial and patellar cartilage. Arthritis Rheum. 52, 3918–3927. Ding, C., Cicuttini, F., Scott, F., Cooley, H., Boon, C., Jones, G., 2006. Natural history of knee cartilage defects and factors affecting change. Arch. Intern. Med. 166, 651–658. Dore, D., Winzenberg, Ding C., Otahal, P., Pelletier, J., Martel-Pelletier, J., Cicuttini, F., Jones, G., 2013. The association between objectively measured physical activity and knee structural change using MRI. Ann. Rheum. Dis. 72, 1170–1175. Eckstein, F., Glaser, C., 2004. Measuring cartilage morphology with quantitative magnetic resonance imaging. Semin Musculoskelet. Radiol. 8, 329–353. Eckstein, F., Schnier, M., Haubner, M., Priebsch, J., Glaser, C., Englmeier, K., Reiser, M., 1998. Accuracy of cartilage volume and thickness measurements with magnetic resonance imaging. Clin. Orthop. Rel. Res. 352, 137–148. Eckstein, F., Ateshian, G., Burgkart, R., Burstein, D., Cicuttini, F., Dardzinski, B., Gray, M., Link, T., Majumdar, S., Mosher, T., Peterfy, C., Totterman, S., Waterton, J., Winalski, C., Felson, D., 2006a. Proposal for a nomenclature for magnetic resonance imaging based measures of articular cartilage in osteoarthritis. Osteoarthr. Cartil. 14, 974–983. Eckstein, F., Hudelmaier, M., Putz, R., 2006b. The effects of exercise on human articular cartilage. J. Anat. 208, 491–512. Felson, D., Lawrence, R., Dieppe, P., Hirsch, R., Helmick, C., Jordan, J., Kington, R., Lane, N., Nevitt, M., Zhang, Y., Sowers, M., McAlindon, T., Spector, T., Poole, R., Yanovski, S., Ateshian, G., Sharma, L., Buckwalter, J., Brandt, K., Fries, J., 2000. Osteoarthritis: new insights: Part 1: the disease and its risk factors. Ann. Intern. Med. 133, 635–646. Gardiner, P., Clark, B., Healy, G., Eakin, E., Winkler, E., Owen, N., 2011. Measuring older adults' sedentary time: reliability, validity and responsiveness. Med. Sci. Sports Exerc. 43, 2127–2133. Hunter, D., Lohmander, L., Makovey, J., Tamez-Pena, J., Totterman, S., Schreyer, E., Frobell, R., 2014. The effect of anterior cruciate ligament injury on bone curvature: exploratory analysis in the KANON trial. Osteoarthr. Cartil. 22, 959–968. Inglis, D., Pui, M., Ioannidis, G., Beattie, K., Boulos, P., Adachi, J., Webber, C., Eckstein, F., 2007. Accuracy and test-retest precision of quantitative cartilage morphology on a 1.0T peripheral magnetic resonance imaging system. Osteoarthr. Cartil. 15, 110–115. Jackson, B., Teichtahl, A., Morris, M., Wluka, A., Davis, S., Cicuttini, F., 2004. The effect of the knee adduction moment on tibial cartilage volume and bone size in healthy women. Rheumatology 43, 311–314. Kean, C., Hinman, R., Bowles, K., Cicuttini, F., Davies-Tuck, M., Bennell, K., 2012. Comparison of peak knee adduction moment and knee adduction moment impulse in distinguishing between severities of knee osteoarthritis. Clin. Biomech. 27, 520–523. Kellgren, J., Lawrence, J., 1957. Radiological assessment of osteoarthritis. Ann. Rheum. Dis. 16, 494–501. Kochersberger, G., McConnell, E., Kuchibhatla, M., Pieper, C., 1996. The reliability, validity and stability of a measure of physical activity in the elderly. Arch. Phys. Med. Rehabil. 77, 793–795. Koo, S., Andriacchi, T., 2007. A comparison of the influence of global functional loads versus local contact anatomy on articular cartilage thickness at the knee. J. Biomech. 40, 2961–2966. Koo, S., Rylander, J., Andriacchi, T., 2011. Knee joint kinematics during walking influences the spatial cartilage thickness distribution in the knee. J. Biomech. 44, 1405–1409. Kothari, M., Guermazi, A., vonIngersleben, G., Miaux, Y., Sieffert, M., Block, J., Stevens, R., Peterfy, C., 2004. Fixed-flexion radiography of the knee provides reproducible joint space width measurements in osteoarthritis. Eur. Radiol. 14, 1568–1573. Maly, M., Robbins, S., Stratford, P., Birmingham, T., Callaghan, J., 2012. Cumulative knee adductor load distinguishes between healthy and osteoarthritis knees—a proof of principle study. Gait Posture S0966-6362 (12). http://dx.doi.org/ 10.1016/j.gaitpost.2012.08.013 00318-9. Miyazaki, T., Wada, M., Kawahara, H., Sato, M., Baba, H., Shimada, S., 2002. Dynamic load at baseline can predict radiographic disease progression in medial compartment knee osteoarthritis. Ann. Rheum. Dis. 61, 617–622. Pearle, A., Warren, R., Rodeo, S., 2005. Basic science of articular cartilage and osteoarthritis. Clin. Sports Med. 24, 1–12. Robbins, S., Birmingham, T., Callaghan, J., Jones, G., Chesworth, B., Maly, M., 2011a. Association of pain with frequency and magnitude of knee loading in knee osteoarthritis. Arthritis Care Res. 63, 991–997.
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Robbins, S., Birmingham, T., Jones, G., Callaghan, J., Maly, M., 2009. Developing an estimate of daily cumulative loading for the knee: examining test–retest reliability. Gait Posture 30, 497–501. Robbins, S., Birmingham, T., Maly, M., Chesworth, B., Giffin, J., 2011b. Comparative diagnostic accuracy of knee adduction moments in knee osteoarthritis: a case for not normalizing to body size. J. Biomech. 44, 968–971. Robertson, D., Dowling, J., 2003. Design and responses of Butterworth and critically damped digital filters. J. Electromyogr. Kinesiol. 13, 569–573. Schache, A., Baker, R., 2007. On the expression of joint moments during gait. Gait Posture 25, 440–452. Schneider, E., Nevitt, M., McCulloch, C., Cicuttini, F., Duryea, J., Eckstein, F., TamezPena, J., 2012. Equivalence and precision of knee cartilage morphometry between different segmentation teams, cartilage regions, and MR acquisitions. Osteoarthr. Cartil. 20, 869–879. Shive, M., Restrepo, A., Totterman, S., Tamez-Pena, J., Schreyer, E., Steinwachs, M., Stanish, W., 2014. Quantitative 3D MRI reveals limited intra-lesional bony overgrowth at 1 year after microfracture-based cartilage repair. Osteoarthr. Cartil. 22, 800–804. Song, J., Semanik, P., Sharma, L., Chang, R., Hochberg, M., Mysiw, W., Bathon, J., Eaton, C., Jackson, R., Kwoh, C., Nevitt, M., Dunlop, D., 2010. Assessing physical activity in persons with knee osteoarthritis using accelerometers: Data from the Osteoarthritis Initiative. Arthritis Care Res. 62, 1724–1732. Stehling, C., Lane, N., Nevitt, M., Lynch, J., McCulloch, C., Link, T., 2010. Subjects with physical activity levels have more severe focal knee lesions diagnosed with 3T MRI: analysis of a non-symptomatic cohort of the osteoarthritis initiative. Osteoarthr. Cartil. 18, 776–786. Tamez-Pena, J., Farber, J., Gonzalez, P., Schreyer, E., Schneider, E., Totterman, S., 2012. Unsupervised segmentation and quantification of anatomical knee
7
features: data from the osteoarthritis initiative. IEEE Trans. Biomed. Eng. 59, 1177–1186. Thorp, L., Sumner, R., Block, J., Moisio, K., Shott, S., Wimmer, M., 2006. Knee joint loading differs in individuals with mild compared with moderate medial knee osteoarthritis. Arthritis Rheum. 54, 3840–3842. Thorp, L., Sumner, R., Wimmer, M., Block, J., 2007. Relationship between pain and medial knee joint loading in mild radiographic knee osteoarthritis. Arthritis Care Res. 57, 1254–1260. Trost, S., McIver, K., Pate, R., 2005. Conducting accelerometer-based activity assessments in field-based research. Med. Sci. Sports Exerc. 37, S531–543. Urquhart, D., Soufan, C., Teichtahl, A., Wluka, A., Hanna, F., Cicuttini, F., 2008. Factors that may mediate the relationship between physical activity and the risk for developing knee osteoarthritis. Arthritis Res. Ther. 10, 203. Vanwanseele, B., Eckstein, F., Smith, R., Lange, A., Foroughi, N., Baker, M., Shnier, R., Fiatarone Singh, M., 2010. The relationship between knee adduction moment and cartilage and meniscus morphology in women with osteoarthritis. Osteoarthr. Cartil. 18, 894–901. Vignon, E., VAlat, J., Rossignol, M., Avouac, B., Rozenberg, S., Thoumie, P., Avouac, J., Nordin, M., Hilliquin, P., 2006. Osteoarthritis of the knee and hip and activity: a systematic international review and synthesis (OASIS). Jt. Bone Spine 73, 442–455. Winter, D., 1990. Biomechanics and Motor Control of Human Movement. WileyInterscience Publication, New York. Wu, G., Cavanagh, P., 1995. ISB recommendations for standardization in the reporting of kinematic data. J. Biomech. 28, 1257–1261.
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis Monica R. Maly a,n, Stacey M. Acker b, Saara Totterman c, José Tamez-Peña c,d, Paul W. Stratford a, Jack P. Callaghan b, Jonathan D. Adachi e, Karen A. Beattie e a
School of Rehabilitation Sciences, McMaster University, Hamilton, Canada Department of Kinesiology, University of Waterloo, Waterloo, Canada c Qmetrics Technologies, Rochester, NY, USA d Escuela de Medicina, Tecnológico de Monterrey, Monterrey, NL, México e Department of Medicine, McMaster University, Hamilton , Canada b
art ic l e i nf o
a b s t r a c t
Article history: Accepted 27 April 2015
The objective was to determine the extent to which the external peak knee adduction moment (KAM) and cumulative knee adductor load explained variation in medial cartilage morphology of the tibia and femur in knee osteoarthritis (OA). Sixty-two adults with clinical knee OA participated (61.576.2 years). To determine KAM, inverse dynamics was applied to motion and force data of walking. Cumulative knee adductor load reflected KAM impulse and loading frequency. Loading frequency was captured from an accelerometer. Magnetic resonance imaging scans were acquired with a coronal fat-saturated sequence using a 1.0 T peripheral scanner. Scans were segmented for medial cartilage volume, surface area of the bone–cartilage interface, and thickness. Forward linear regressions assessed the relationship of loading variables with cartilage morphology unadjusted, then adjusted for covariates. In the medial tibia, age and peak KAM explained 20.5% of variance in mean cartilage thickness (p o0.001). Peak KAM alone explained 12.3% of the 5th percentile of medial tibial cartilage thickness (i.e., thinnest cartilage region) (p ¼0.003). In the medial femur, sex, BMI, age, and peak KAM explained 44% of variance in mean cartilage thickness, with peak KAM contributing 7.9% (po0.001). 20.7% of variance in the 5th percentile of medial femoral cartilage thickness was explained by BMI and peak KAM (p¼ 0.001). In these models, older age, female sex, greater BMI, and greater peak KAM related with thinner cartilage. Models of KAM impulse produced similar results. In knee OA, KAM peak and impulse, but not loading frequency, were associated with cartilage thickness of the medial tibia and femur. & 2015 Elsevier Ltd. All rights reserved.
Keywords: Arthritis Cartilage Articular Knee joint Locomotion
1. Introduction Joint loading is implicated in knee osteoarthritis (OA) progression (Andriacchi et al., 2009). Over six years, an increase by 25% of the peak knee adduction moment (KAM) corresponded with a 6.46 times greater risk for progressive joint space narrowing in knee OA (Miyazaki et al., 2002). Similar findings have been produced with magnetic resonance imaging (MRI). In 180 people with medial knee OA, more severe cartilage defects were associated with higher peak KAM (Creaby et al., 2010). Five-year changes in medial tibial cartilage thickness were predicted by baseline peak KAM and peak knee flexion moment in 16 people with knee OA (Chehab et al., 2014). However, the peak KAM does n Correspondence to: McMaster University, Room 435 IAHS, 1400 Main St. West, Hamilton, ON, Canada L8S 1C7. Tel.: þ 1 905 525 9140x27823; fax: þ1 905 524 0069. E-mail address: [email protected] (M.R. Maly).
not always relate with cartilage damage. Peak KAM was unrelated to progression of cartilage defects over 12 months in medial knee OA (Bennell et al., 2011). It may be unreasonable to expect that a laboratory measure captured during a single instant would share a relationship with cartilage features that reflect long-term loading. The magnitude, duration and repetition of weight-bearing activities dictate the responses of cartilage to mechanical loads (Pearle et al., 2005; Maly et al., 2012). Compared to peak KAM, the combination of KAM impulse and loading frequency encountered in daily activity may better reflect cartilage morphology. Accounting for both KAM impulse and loading frequency together explained variance in pain among adults with knee OA (Robbins et al., 2011a, 2011b) and was superior to the peak KAM in distinguishing between adults with and without knee OA (Maly et al., 2012). KAM impulse reflects the total duration and magnitude of medial knee loading during one stride. KAM impulse was associated with pain intensity in knee OA (Thorp et al., 2007; Robbins et al., 2011a, 2011b). In radiographic knee OA, this moment-time
http://dx.doi.org/10.1016/j.jbiomech.2015.04.039 0021-9290/& 2015 Elsevier Ltd. All rights reserved.
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
integral distinguished between moderate and mild radiographic disease; whereas the peak KAM was not different between groups (Thorp et al., 2006). A higher KAM impulse, but not peak KAM, was associated with greater loss of medial tibial cartilage volume over 12 months in knee OA (Bennell et al., 2011). Loading frequency is also important in cartilage degradation. Epidemiological studies of sport and occupation show a dose– response relationship between loading frequency and knee OA incidence (Coggon et al., 2000; Vignon et al., 2006; Stehling et al., 2010). A 10-year follow-up of 16,961 patients demonstrated that high loading frequency, noted by distance of a walk or run, had an odds ratio of 2.4 (95%CI 1.5–3.9) for symptomatic knee OA in men (Cheng et al., 2000). Greater steps/day increased cartilage loss over 2.7 years in those with low cartilage volume at baseline; but interestingly protected against cartilage loss in those with greater cartilage volume at baseline (p ¼0.046) (Dore et al., 2013). High loading frequency (4 10,000 steps/day) was associated with a greater severity of cartilage defects in this sample (Dore et al., 2013). The purpose of this study was to determine the extent to which the peak KAM, and cumulative knee adductor load (KAM impulse, loading frequency) explained variation in medial tibial and femoral cartilage morphology in knee OA. We hypothesized that peak KAM would relate inversely with the 5th percentile of cartilage thickness; while elements of cumulative load would show stronger relationships with morphology. Specifically, we hypothesized that KAM impulse would relate inversely with cartilage volume and thickness, and positively with surface area of the bone–cartilage interface; while loading frequency would relate inversely with cartilage volume. 2. Methods This cross-sectional study was approved by the Institutional Human Research Ethics Board.
Table 1 Nomenclature for measures of cartilage morphology are consistent with published norms (Eckstein et al., 2006a). The medial tibia and medial femur were represented by the prefixes MT and MF respectively. VC refers to volume of cartilage. tAB refers to total area of subchondral bone. cAB refers to area of subchondral bone covered in cartilage. Finally, ThCtAB refers to cartilage thickness over total subchondral bone area. All measures represent mean values from the medial compartment of the knee.
Cartilage volume (mm3) Surface area of bone–cartilage interface (mm2) Cartilage thickness (mm)
Tibial
Femoral
MT.VCtAB MT.cAB MT.ThCtAB
MF.VCtAB MF.cAB MF.ThCtAB
at the femoral trochlea in images of OA knees obtained with a 3.0 T MRI (TamezPena et al., 2012) and has evaluated cartilage repair (Shive et al., 2014) and bone morphology (Hunter et al., 2014). The accuracy of segmentation from scans acquired with a 1.5 T magnet produced data within 4–6% of values from a 3.0 T (Schneider et al., 2012). No statistical differences in cartilage measurements existed between 1.5 T and 1.0 T magnets (Inglis et al., 2007). Briefly, eight atlases were used to segment each scan. Each atlas was morphed to match anatomy using a spline deformation with a mutual information similarity metric. After spline morphing, the edges of the segmented image were deformed using a free-form registration algorithm. The morphed cartilage segmentation was then corrected by outlier detection: voxels whose signal was three standard deviations higher than average were classified as outliers. Segmentations from the eight atlases were fused using a fuzzy voting algorithm, and statistically relaxed to create the final segmentation (Tamez-Pena et al., 2012). All segmentations were reviewed by an experienced radiologist (ST) for quality. The following were calculated from the medial tibia and femur: volume (VCtAB; mm3), surface area of the bone–cartilage interface (cAB; mm2), mean thickness (ThCtAB; mm) and mean of the 5th percentile thickness (mm). These acronyms (Table 1) are consistent with established nomenclature (Eckstein et al., 2006a, 2006b). Coronal weight-bearing knee radiographs were obtained in a standardized fixed-flexion position using a SynaflexerTM (Kothari et al., 2004). This frame places the feet in 5° of external rotation and 20° of knee flexion. Each digital radiograph was evaluated by one experienced radiologist (ST) to yield Kellgren and Lawrence (K–L) scores (Kellgren and Lawrence, 1957) and frontal plane alignment (anatomical axis). 2.3. Biomechanical Assessment
2.1. Participants Community-dwelling adults between 40 and 70 years who met the American College of Rheumatology (ACR) clinical criteria for knee OA were recruited from rheumatologist's and orthopaedic surgeon's offices. These guidelines include having knee pain on most days of the month and at least three of the following: Z50 years of age; stiffness o30 min; crepitus; bony tenderness; bony enlargement; no palpable warmth (Altman et al., 1986). Exclusion criteria included other forms of arthritis; non-arthritic disease; intra-articular therapies; or previous knee surgeries (e.g. osteotomy, replacement, partial/complete menisectomy, ligament reconstruction). Potential participants were excluded if they required an adaptive walking aid; sustained leg trauma within the past 3 months; had ipsilateral hip or ankle conditions; or had contraindications to MRI. In those with bilateral knee OA, the knee with more severe symptoms was designated as the study knee. All participants provided written, informed consent. 2.2. Cartilage morphometry Each participant underwent an MRI scan of the study knee using a 1.0 T peripheral MRI scanner (GE Healthcare, USA). Participants were seated with the knee fully extended and centered in the iso-center of the 180 mm removable quadrature volume transmit-receive coil. Padding around the knee, thigh, and leg limited movement. Sagittal gradient-echo and axial fast spin-echo localizer scans were performed (2–3 min). A coronal fat-saturated spoiled gradient recalled acquisition in the steady-state (SPGR) was acquired: TR 60 ms; TE 12.4 ms (or minimum); flip angle 40°; bandwidth 30 kHz; matrix 512 256 (frequency phase); 1 excitation; field of view 150 mm; slice thickness 1.5 mm; 56–64 partitions depending on patient size. Scan time was 15–16 min. MRI fat-saturated sequences (i.e., SPGR) from similar lower level magnets quantify cartilage morphology with accuracy of 13% to 3% and precision around 4.0% (Eckstein et al., 1998; Stammberger et al., 1999; Burgkart et al. 2001; Eckstein and Glaser, 2004). Medial tibial and femoral cartilage morphology was segmented from these images using a highly automated, atlas-based method (Qmetrics, Rochester, NY, USA) (Tamez-Pena et al., 2012). This method yielded test–retest precision of cartilage thickness values between 0.014 mm (0.6%) at the femur and 0.038 mm (1.6%)
Within one week of the MRI, a laboratory visit was conducted. To calculate the peak and impulse of the external KAM, gait analyses were conducted during barefoot walking at a self-selected speed. Kinetics were collected using a synchronized floor-mounted force plate (OR6-7, Advanced Mechanical Technology Inc., Watertown, MA, USA) sampled at 1000 Hz. Concurrently, kinematics were collected using three Optotrak Certus banks (nine cameras) (Northern Digital Inc., Waterloo, ON, Canada) sampled at 100 Hz. Infrared emitting diodes, arranged on rigid bodies in clusters of three, were secured to the lateral sacrum, thigh, shank and foot. The pelvis was digitized using six landmarks (bilateral anterior and posterior superior iliac spines, greater trochanters). Virtual markers were also created for the greater trochanter, medial and lateral femoral and tibial condyles, tibial tuberosity, fibular head, medial and lateral malleoli, calcaneus, and first, second and fifth metatarsal heads. A static reference trial was recorded to determine neutral pelvic and lowerlimb joint angles. Participants ambulated until five trials were captured where the participant struck the force plate cleanly with the study leg. Using commercial software (Visual 3D, C-Motion, Inc., Germantown, MD, USA), marker data were dual-passed through a low-pass Butterworth filter with a 6 Hz cut-off frequency (Robertson and Dowling 2003). Inverse dynamics was used to calculate external knee moments (Winter 1990) in a three-dimensional floating axis coordinate system where flexion/extension occurred about the lateral-medial axis of the thigh, internal/external rotation occurred about the distal-proximal axis of the shank, and adduction/abduction occurred about a floating axis (intermediate axis perpendicular to both lateral-medial and distal-proximal axes) (Cole et al. 1993; Wu and Cavanagh 1995; Schache and Baker 2007). The rotation sequence was consistent with a Cardan XYZ rotation (Wu and Cavanagh 1995). The KAM peak and impulse are highly reliable, stable measures in knee OA (Birmingham et al. 2007; Robbins et al. 2009). Peak KAM was reported in Nm/kg (Winter 1990). Mean KAM impulse of five trials was calculated using the trapezoid rule, which integrated only positive values from stance (Matlab 7.0.1, Natick, MA, USA). Our interest focused on the medial knee; thus negative (abduction) values were not subtracted from the adduction impulse because these values do not subtract from the loading experienced by the medial knee. To reflect the theoretical underpinnings of cumulative load, which represents the total load exposure experienced by tissues for a given time period, KAM impulse values were in non-normalized units of Nm●s. Non-normalization ensures that the impulse
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ reflects the absolute cumulative knee load during stride (Browning and Kram 2007). Loading frequency was the daily average number of steps on the study leg using a tri-axial accelerometer (GTX3 þ, ActiGraph, Pensacola, Florida, USA) (Kochersberger et al. 1996; Gardiner et al. 2011). The sampling rate was 30 Hz. Participants wore the accelerometer for waking hours, except for activities involving water, for seven consecutive days. The accelerometer was secured on a waist-belt, directly above the thigh midline of the study leg. Participants completed an activity logbook to record periods of non-wear. Any day where the accelerometer was not worn for least 10 hours was excluded (Song et al. 2010). Accelerometry in elderly samples required five days to estimate daily physical activity (Trost et al. 2005). Step counts from the first five full days were averaged. 2.4. Data analyses A series of analyses were completed to build linear models that expressed the relationship of mechanics with medial cartilage morphology. First, forward linear regressions were computed where measures of medial cartilage morphology were the dependent variables: volume, surface area of the bone-cartilage interface, mean thickness, and 5 th percentile thickness of the tibia and femur. Two models of independent variables were composed. In Model 1, the relationship of peak KAM with tibial and femoral cartilage morphology was explored. In Model 2, relationships of cartilage morphology with KAM impulse and steps/day were explored together, representing cumulative knee adductor load (Robbins et al. 2011a; 2011b; Maly et al. 2012). These models were explored unadjusted. Next, adjusted forward regressions were computed using covariates of age, sex, and BMI, provided these were significant (i.e., p o0.05). Older age, female sex and greater BMI were associated with the development and progression of cartilage defects over two years (Ding et al. 2006) and greater cartilage loss was observed in women than men, particularly with aging (Ding et al. 2007). Finally, if the independent variables of interest (e.g., peak KAM, KAM impulse, loading frequency) were not entered in the adjusted model, we conducted a follow-up analysis that included statistically significant covariates and independent variables of interest to clearly show that the independent variables of interest did not contribute to the model. The probability of F statistic was set at 0.05 to enter. Statistical significance was set at P o 0.05. Histograms of the residuals and plots of the residuals against the predicted values of the dependent variables were utilized to examine the appropriateness of the regression analysis. Tolerance (Z 0.908) and variance inflation factors ( r 1.1) demonstrated that multicollinearity was not present. Analyses were completed using SPSS 21.0 (SPSS Inc., Chicago, USA).
3. Results Sixty-two adults (11 men) participated (Table 2). This group included 27, 18 and 17 participants with K-L scores of 2, 3 and 4, respectively. All had radiographic medial OA. Anatomical alignment was 3.4 72.6° varus (range¼-5.9 to 11.3°) with 19 of 62 participants demonstrating valgus (alignment o0°). Table 2 Descriptive statistics including demographics, gait mechanics and cartilage morphology (n¼ 62). Variable
Mean (Standard Deviation) Minimum– maximum
Age (y) Body Mass (kg) Height (m) Body Mass Index (kg/m2) Peak KAM (Nm/kg) KAM Impulse (Nm●s) Load Frequency (steps/day)a MT.VCtAB (mm3) MT.cAB (mm2) MT.ThCtAB (mm) Tibial 5th% Thickness (mm) MF.VCtAB (mm3) MF.cAB (mm2) MF.ThCtAB (mm) Femoral 5th% Thickness (mm)
61.5 (6.2) 74.1 (16.0) 1.61 (0.10) 28.2 (5.7) 0.35 (0.17) 9.14 (6.29) 3898 (1951) 1652 (384) 918 (148) 1.8 (0.2) 0.8 (0.1) 1740 (398) 877 (114) 1.8 (0.2) 0.8 (0.1)
41–70 49.4–117.0 1.07–1.85 19.7–41.8 0.01–0.72 0.87–26.94 644–10968 934–2929 591–1463 1.3–2.3 0.5–1.0 1201–3225 657–1188 1.5–2.5 0.6–0.9
KAM is the external knee adduction moment. a
Steps/day reflect loading frequency for the study leg only.
3
3.1. Tibial cartilage morphology Table 3 summarizes the models for the medial tibia. Loading frequency shared a positive relationship with MT.VCtAB; adjusting for sex and age eliminated this relationship. Both KAM impulse and loading frequency were positively related to MT.cAB. However adjusting for demographics eliminated these from the final model. Thus, male sex was associated with larger cartilage volume and surface area; and young age related to greater medial tibial cartilage volume. Cartilage thickness was inversely related with KAM. Peak KAM explained 6.6% of the variance in MT.ThCtAB after adjusting for age (p o0.001). The KAM impulse explained 5.1% of the variance in MT.ThCtAB after adjusting for age and sex (p o0.001). Beta coefficients for the adjusted peak KAM, in Model 1, and KAM impulse, in Model 2, were not appreciably different from the unadjusted values. Peak KAM alone explained 12.3% of the variance in the mean 5th percentile of cartilage thickness in the medial tibia (p ¼0.003), such that a greater peak KAM related with thinner cartilage. KAM impulse explained 15.6% of variance in the mean 5th percentile of cartilage thickness in the medial tibia (p o0.005). Loading frequency did not contribute to variance in tibial cartilage morphology in Model 2 after adjusting for age, sex and BMI. 3.2. Femoral cartilage morphology Table 4 summarizes the models for the medial femur. Male sex was associated with larger MF.VCtAB and a larger MF.cAB; and young age related to greater MF.VCtAB. Peak KAM explained 7.9% of the variance in MF.ThCtAB, where a greater peak KAM related with thinner cartilage, after adjusting for sex, BMI and age (p o0.001). Similarly, 8.4% of the variance in MF.ThCtAB was explained by the KAM impulse after adjusting for sex, BMI and age (p o0.001). BMI and the peak KAM explained 20.7% of the variance in the mean 5th percentile of cartilage thickness in the medial femur, with peak KAM contributing 12.5% (p o0.001). Both BMI and peak KAM inversely related with mean 5th percentile of cartilage thickness. KAM impulse explained 10.5% of variance in the mean 5th percentile of medial femoral cartilage thickness after adjusting for BMI (p o0.001). Beta coefficients for the unadjusted and adjusted peak KAM and KAM impulse were similar. Loading frequency did not contribute to variance in any of the femoral cartilage morphology measures in Model 2.
4. Discussion In clinical, radiographic knee OA, the peak KAM and KAM impulse explained variance in cartilage thickness of the medial tibia and femur. Greater medial knee loading was associated with thinner cartilage. No relationship existed between loading frequency and cartilage morphology. These findings highlight the importance of biomechanics in cartilage degradation by demonstrating that medial knee loads are associated with cartilage morphometry in vivo. A novel contribution is exploring loading frequency in the relationships of biomechanics with cartilage morphology in knee OA. Loading frequency was not related to cartilage morphology after accounting for demographics. Thus, the cumulative knee adductor load approach was not useful. A cross-sectional snapshot of loading frequency may not reflect an individual’s lifespan of physical activity. Previous work using a variety of methods to characterize loading frequency and knee health have yielded conflicting results (Urquhart et al., 2008). Mediators of this relationship include sex, age, BMI, body composition, strength, injury, and gait biomechanics (Urquhart et al., 2008). People who have
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
Table 3 Forward linear regression models of medial tibial cartilage morphology. Model 1 includes the peak knee adduction moment (KAM) as a potential predictor. Model 2 includes the KAM impulse and loading frequency as potential predictors. The models are presented unadjusted and adjusted for age, sex and BMI. The final model is presented in bold. Dependent variable
Model 1 Predictors
MT.VCtAB (mm3)
MT.cAB (mm2)
MT.ThCtAB (mm)
Unadjusted Peak KAM
Adjusted R2
Unstandardized β coefficient
P
0.01
232.936
0.429
Adjusted Sex þ Age Sex Age Peak KAM
0.444 0.502 0.509
643.957 15.764 272.254
0.001 0.001 0.007 0.186
Unadjusted Peak KAM
0.02
29.862
0.793
Adjusted Sex Peak KAM
0.499 0.488
280.746 4.540
0.001 0.955
Unadjusted Peak KAM
0.068
0.351
0.023
Adjusted Age þ Peak KAM Age Peak KAM
5th Percentile tibial thickness (mm)
Model 2
Unadjusted Peak KAM
Adjusted Peak KAM
0.001 0.139 0.205
0.123
0.123
0.011 0.325
0.205
0.205
0.003 0.019
0.003
0.003
greater step counts may possess other features, such as greater fitness, that may protect cartilage health but were not captured in this study. Baseline knee health also appears important. Among 405 adults between 51 and 81 years of age, high frequency loading protected against cartilage loss over 2.7 years in those with large cartilage volumes at baseline; but facilitated cartilage loss in those with smaller cartilage volume at baseline (Dore et al., 2013). Mechanical loads of magnitudes beyond the physiological capacity of cartilage facilitate degeneration of this tissue (Felson et al., 2000; Miyazaki et al., 2002; Urquhart et al., 2008; Bennell et al., 2011). Our findings identify that large KAMs during walking were associated with the presence of thinned tibiofemoral cartilage. While sex, age and BMI were associated with cartilage morphology in some models, the consistency of the beta coefficients for KAM peak and impulse in the unadjusted and adjusted models suggests that these covariates did not alter the association between medial loading and cartilage thickness of the medial tibia and femur. Previous work has produced conflicting results. While a high KAM peak and impulse related to meniscal extrusion (p o0.001) and lower medial meniscus height (p¼ 0.010), the
Predictors
Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Sex þ Age Sex Age KAM impulse Loading frequency Unadjusted KAM impulse þLoading frequency KAM impulse Loading frequency Adjusted Sex KAM impulse Loading frequency Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Age þ Sex þ KAM impulse
Adjusted R2
Unstandardized β coefficient
P
0.018 0.013 0.104
11.436 0.065
0.152 0.012
0.444 0.502 0.493 0.499
643.957 15.764 0.696 0.025
0.001 0.001 0.003 0.911 0.210
0.004 0.085 0.154
7.826 0.022
0.010 0.021
0.499 0.497 0.513
280.746 2.794 0.012
0.001 0.242 0.102
0.080 0.006 0.055
0.005 0.001
0.256 0.055 0.001
Age Sex
0.150 0.202
0.011 0.175
0.006 0.009
KAM impulse Loading frequency
0.253 0.245
0.008 0.001
0.035 0.510
0.156
0.006
0.005
0.129
0.001
0.211
0.156 0.129
0.006 0.001
0.005 0.211
Unadjusted KAM impulse Loading frequency Adjusted KAM impulse Loading frequency
thickness of cartilage in the tibia and femur were unrelated to the KAM peak and impulse in 45 women with clinical knee OA (Vanwanseele et al., 2010). Another study conducted cross-sectional analyses of 180 people with radiographic knee OA, which showed that both the peak KAM and KAM impulse were positively associated with medial knee cartilage defect scores (Creaby et al., 2010). Sample size and differences in measurement (e.g. cartilage thickness from segmentation versus categorization of cartilage defects) may be responsible for this discrepancy. The 5th percentile of cartilage thickness quantified the average thickness of thinnest cartilage tissue. This measure likely equates with previous categorizations of cartilage defects (Creaby et al., 2010; Bennell et al., 2011). Several methodological differences exist between Creaby’s work (Creaby et al., 2010) and the present study. Creaby et al. (2010) recruited a narrower range of radiographic severities of knee OA (mild–moderate), analyzed shod gait, and utilized manual segmentation from MRI scans from two different 1.5T whole-body scanners. The current study recruited people with clinical knee OA (K–L 2–4), analyzed barefoot gait and used highlyautomated segmentation of scans from one 1.0 T peripheral
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Table 4 Forward linear regression models of medial femoral cartilage morphology. Model 1 includes the peak knee adduction moment (KAM) as a potential predictor. Model 2 includes the KAM impulse and loading frequency as potential predictors. The models are presented unadjusted and adjusted for age, sex and BMI. The final model is presented in bold. Dependent variable
Model 1 Predictors
MF.VCtAB (mm3)
MF.cAB (mm2)
MF.ThCtAB (mm)
5th Percentile femoral thickness (mm)
Unadjusted Peak KAM
Model 2 Adjusted R2
Unstandardized β coefficient
P
0.005
329.384
0.257
Adjusted Sex þ Age Sex Age Peak KAM
0.415 0.454 0.473
628.784 13.360 370.354
0.001 0.001 0.027 0.085
Unadjusted Peak KAM
0.017
4.715
0.957
Adjusted Sex Peak KAM
0.439 0.433
205.643 20.409
0.001 0.756
Unadjusted Peak KAM
0.077
0.322
0.017
Adjusted Sex þ BMI þ Age þ Peak KAM Sex BMI Age Peak KAM
0.170 0.269 0.362 0.441
0.191 0.011 0.008 0.310
0.001 0.001 0.006 0.004
Unadjusted Peak KAM
0.124
0.126
0.003
Adjusted BMI þ Peak KAM BMI Peak KAM
0.001
0.001 0.082 0.207
0.003 0.125
scanner. The consistency in findings despite methodological differences provides confidence that the relationships between walking mechanics and cartilage defects exist. The peak KAM, which lasts a fraction of one second during stride, and the KAM impulse, which reflects both the duration and magnitude of medial loads throughout stride, produced similar results. (Vanwanseele et al., 2010) noted a similar pattern in 45 women with clinical knee OA. The distinction between KAM peak and impulse may not be important to cartilage morphology. Instead, a general picture of load magnitude appeared sufficient to relate with cartilage thickness. The peak KAM was less sensitive than the KAM impulse when distinguishing between radiographic severities of knee OA (Thorp et al., 2006; Robbins et al., 2011a; 2011b; Kean et al., 2012). Cross-sectionally, both the peak KAM and KAM impulse were positively associated with severity of cartilage defects (Creaby et al., 2010). However, follow-up on the
0.011 0.002
Predictors
Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Sex þ Age Sex Age KAM impulse Loading frequency Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Sex KAM impulse Loading frequency
Adjusted R2
P
0.201 0.018 0.022
1.330 0.047
0.870 0.075
0.413 0.452 0.477 0.469
630.332 13.646 11.887 0.008
0.001 0.001 0.031 0.069 0.699
0.064 0.050 0.074
4.864 0.012
0.044 0.124
0.440 0.442 0.436
207.039 0.971 0.004
0.001 0.621 0.514
Unadjusted KAM impulse þ Loading frequency KAM impulse Loading frequency Adjusted Sex þ BMI þ Age þ KAM impulse
0.053 0.095
Sex BMI Age KAM impulse Loading frequency
0.170 0.263 0.360 0.444 0.435
0.024 0.008 0.001
0.045 0.062 0.001
Unadjusted KAM impulse þ Loading frequency KAM impulse 0.164 Loading frequency 0.149 Adjusted BMI þ KAM impulse BMI KAM impulse Loading frequency
Unstandardized β coefficient
0.107 0.212 0.183
0.237 0.008 0.008 0.010 0.001
0.001 0.035 0.012 0.004 0.701
0.004 0.004 0.001
0.001 0.880 0.001
0.003 0.003 0.001
0.051 0.006 0.791
same sample showed that over one year, only KAM impulse was related to cartilage volume loss (Bennell et al., 2011). It is possible that changes in cartilage volume over one year were too small to observe this association. Surface area of the bone–cartilage interface was unrelated to biomechanical measures, in contrast to previous work. Creaby et al. (2010) showed that loading may have a role in widening and flattening the tibial plateau. Large mechanical loads attributed to training (triathletes) and obesity related to an expansion of tibial surface area (Ding et al., 2005). Subchondral bone area was directly related to the peak KAM (Jackson et al., 2004; Dempsey et al., 2013) and related to the ratio of medial to lateral tibial subchondral bone area (Vanwanseele et al., 2010). The discrepancy between findings in the current study and previous work may reflect definitions characterizing surface area. We quantified areas of bone that were covered in cartilage; whereas previous work
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
identified the subchondral bone surface area, regardless of the presence of cartilage (Jackson et al., 2004; Vanwanseele et al., 2010). Finally, cartilage volumes of the medial tibia and femur were unrelated to the biomechanical measures. It is likely that regionspecific analyses, such as those proposed by Eckstein et al. (2006a, 2006b) are necessary. Regional variations in knee cartilage thickness and volume are associated with loading patterns during gait (Koo and Andriacchi, 2007). High loads at the knee occur at heelstrike where the knee is near full extension (Koo et al., 2011) when the thickest regions of femoral and tibial cartilage are aligned (Andriacchi et al., 2004; Eckstein et al., 2006a; 2006b). The cross-sectional nature of this study limits the conclusions to associations only. While all participants had radiographic evidence of medial disease, we did not exclude those with concurrent lateral disease. Also, the anatomical axis angles for this sample suggest that participants were either varus, or demonstrated slight valgus. Together these data suggest that the participants likely loaded the medial knee to a greater extent than the lateral. Finally, eight dependent variables were tested. Multiplicity may have increased the likelihood of identifying statistically significant relationships. 4.1. Conflicts of interest statement Saara Totterman and José Tamez-Peña are employed by QMetrics Technologies.
Acknowledgements Funded by Canadian Institutes of Health Research (CIHR) Operating Grant (#102643 MRM). Salary support was provided by a CIHR New Investigator Award (MRM), Tier I Canada Research Chair (JPC) and Arthritis Society/Canadian Arthritis Network Scholar Award (KAB). Equipment required to conduct this research was purchased with the support of the Canadian Foundation for Innovation Leaders Opportunity Fund and the Ministry of Research and Innovation – Ontario Research Fund (MRM).
References Altman, R., Asch, E., Bloch, D., Bole, G., Borenstein, D., Brandt, K., Christy, W., Cooke, T., Greenwald, R., Hochberg, M., Howell, D., Kaplan, D., Koopman, W., Longley, S., Mankin, H., McShane, D., Medsger, T., Meenan, R., Mikkelsen, W., Moskowitz, R., Murphy, W., Rothschild, B., Segal, M., Sokoloff, L., Wolfe, F., 1986. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Arthritis Rheum. 29 (8), 1039–1049. Andriacchi, T., Koo, S., Scanlan, S., 2009. Gait mechanics influence healthy cartilage morphology and osteoarthritis of the knee. J. Bone Jt. Surg. Am. 91, 95–101. Andriacchi, T., Mundermann, A., Smith, R., Alexander, E., Dyrby, C., Koo, S., 2004. A framework for the in vivo pathomechanics of osteoarthritis at the knee. Ann. Biomed. Eng. 32, 447–457. Bennell, K., Bowles, K., Wang, Y., Cicuttini, F., Davies-Tuck, M., Hinman, R., 2011. Higher dynamic medial knee load predicts greater cartilage loss over 12 months in medial knee osteoarthritis. Ann. Rheum. Dis. 70, 1770–1774. Birmingham, T., Hunt, M., Jones, I., Jenkyn, T., Giffin, J., 2007. Test–retest reliability of the peak knee adduction moment during walking in patients with medial compartment knee osteoarthritis. Arthritis Rheum. 57, 1012–1017. Browning, R., Kram, R., 2007. Effect of obesity on the biomechanics of walking at different speeds. Med. Sci. Sports Exerc. 39, 1632–1641. Burgkart, R., Glaser, C., Hyhlik-Durr, A., Englmeier, K., Reiser, M., Eckstein, F., 2001. Magnetic resonance imaging-based assessment of cartilage loss in severe osteoarthritis: Accuracy, precision, and diagnostic value. Arthritis Rheum. 44, 2072–2077. Chehab, E., Favre, J., Erhart-Hledik, J., Andriacchi, T., 2014. Baseline knee adduction and flexion moments during walking are both associated with 5 year changes in patients with medial knee osteoarthritis. Osteoarthr. Cartil. 22, 1833–1839. Cheng, Y., Macera, C., Davis, D., Ainsworth, B., Troped, P., Blair, S., 2000. Physical activity and self-reported, physician-diagnosed osteoarthritis: is physical activity a risk factor? J. Clin. Epidemiol. 53, 315–322.
Coggon, D., Croft, P., Kellingray, S., Barrett, D., McLaren, M., Cooper, C., 2000. Occupational physical activities and osteoarthritis of the knee. Arthritis Rheum. 43, 1443–1449. Cole, G., Nigg, B., Ronsky, J., Yeadon, M., 1993. Application of the joint coordinate system to three-dimensional joint attitude and movement representation: a standardization protocol. J. Biomech. Eng. 115, 344–349. Creaby, M., Wang, Y., Bennell, K., Hinman, R., Metcalf, B., Bowles, K., Cicuttini, F., 2010. Dynamic knee loading is related to cartilage defects and tibial plateau bone area in medial knee osteoarthritis. Osteoarthr. Cartil. 18, 1380–1385. Dempsey, A., Wang, Y., Thorlund, J., Mills, T., Bennell, K., Metcalf, B., Hanna, F., Cicuttini, F., Lloyd, D., 2013. The relationship between patellofemoral and tibiofemoral morphology and gait biomechanics following arthroscopic medial meniscectomy. Knee Surg. Sports Traumatol. Arthrosc. 21, 1097–1103. Ding, C., Cicuttini, F., Blizzard, L., Scott, F., Jones, G., 2007. A longitudinal study of the effect of sex and age on rate of change in knee cartilage volume in adults. Rheumatology 46, 273–279. Ding, C., Cicuttini, F., Scott, F., Boon, C., Jones, G., 2005. Association of prevalent and incident knee cartilage defects with loss of tibial and patellar cartilage. Arthritis Rheum. 52, 3918–3927. Ding, C., Cicuttini, F., Scott, F., Cooley, H., Boon, C., Jones, G., 2006. Natural history of knee cartilage defects and factors affecting change. Arch. Intern. Med. 166, 651–658. Dore, D., Winzenberg, Ding C., Otahal, P., Pelletier, J., Martel-Pelletier, J., Cicuttini, F., Jones, G., 2013. The association between objectively measured physical activity and knee structural change using MRI. Ann. Rheum. Dis. 72, 1170–1175. Eckstein, F., Glaser, C., 2004. Measuring cartilage morphology with quantitative magnetic resonance imaging. Semin Musculoskelet. Radiol. 8, 329–353. Eckstein, F., Schnier, M., Haubner, M., Priebsch, J., Glaser, C., Englmeier, K., Reiser, M., 1998. Accuracy of cartilage volume and thickness measurements with magnetic resonance imaging. Clin. Orthop. Rel. Res. 352, 137–148. Eckstein, F., Ateshian, G., Burgkart, R., Burstein, D., Cicuttini, F., Dardzinski, B., Gray, M., Link, T., Majumdar, S., Mosher, T., Peterfy, C., Totterman, S., Waterton, J., Winalski, C., Felson, D., 2006a. Proposal for a nomenclature for magnetic resonance imaging based measures of articular cartilage in osteoarthritis. Osteoarthr. Cartil. 14, 974–983. Eckstein, F., Hudelmaier, M., Putz, R., 2006b. The effects of exercise on human articular cartilage. J. Anat. 208, 491–512. Felson, D., Lawrence, R., Dieppe, P., Hirsch, R., Helmick, C., Jordan, J., Kington, R., Lane, N., Nevitt, M., Zhang, Y., Sowers, M., McAlindon, T., Spector, T., Poole, R., Yanovski, S., Ateshian, G., Sharma, L., Buckwalter, J., Brandt, K., Fries, J., 2000. Osteoarthritis: new insights: Part 1: the disease and its risk factors. Ann. Intern. Med. 133, 635–646. Gardiner, P., Clark, B., Healy, G., Eakin, E., Winkler, E., Owen, N., 2011. Measuring older adults' sedentary time: reliability, validity and responsiveness. Med. Sci. Sports Exerc. 43, 2127–2133. Hunter, D., Lohmander, L., Makovey, J., Tamez-Pena, J., Totterman, S., Schreyer, E., Frobell, R., 2014. The effect of anterior cruciate ligament injury on bone curvature: exploratory analysis in the KANON trial. Osteoarthr. Cartil. 22, 959–968. Inglis, D., Pui, M., Ioannidis, G., Beattie, K., Boulos, P., Adachi, J., Webber, C., Eckstein, F., 2007. Accuracy and test-retest precision of quantitative cartilage morphology on a 1.0T peripheral magnetic resonance imaging system. Osteoarthr. Cartil. 15, 110–115. Jackson, B., Teichtahl, A., Morris, M., Wluka, A., Davis, S., Cicuttini, F., 2004. The effect of the knee adduction moment on tibial cartilage volume and bone size in healthy women. Rheumatology 43, 311–314. Kean, C., Hinman, R., Bowles, K., Cicuttini, F., Davies-Tuck, M., Bennell, K., 2012. Comparison of peak knee adduction moment and knee adduction moment impulse in distinguishing between severities of knee osteoarthritis. Clin. Biomech. 27, 520–523. Kellgren, J., Lawrence, J., 1957. Radiological assessment of osteoarthritis. Ann. Rheum. Dis. 16, 494–501. Kochersberger, G., McConnell, E., Kuchibhatla, M., Pieper, C., 1996. The reliability, validity and stability of a measure of physical activity in the elderly. Arch. Phys. Med. Rehabil. 77, 793–795. Koo, S., Andriacchi, T., 2007. A comparison of the influence of global functional loads versus local contact anatomy on articular cartilage thickness at the knee. J. Biomech. 40, 2961–2966. Koo, S., Rylander, J., Andriacchi, T., 2011. Knee joint kinematics during walking influences the spatial cartilage thickness distribution in the knee. J. Biomech. 44, 1405–1409. Kothari, M., Guermazi, A., vonIngersleben, G., Miaux, Y., Sieffert, M., Block, J., Stevens, R., Peterfy, C., 2004. Fixed-flexion radiography of the knee provides reproducible joint space width measurements in osteoarthritis. Eur. Radiol. 14, 1568–1573. Maly, M., Robbins, S., Stratford, P., Birmingham, T., Callaghan, J., 2012. Cumulative knee adductor load distinguishes between healthy and osteoarthritis knees—a proof of principle study. Gait Posture S0966-6362 (12). http://dx.doi.org/ 10.1016/j.gaitpost.2012.08.013 00318-9. Miyazaki, T., Wada, M., Kawahara, H., Sato, M., Baba, H., Shimada, S., 2002. Dynamic load at baseline can predict radiographic disease progression in medial compartment knee osteoarthritis. Ann. Rheum. Dis. 61, 617–622. Pearle, A., Warren, R., Rodeo, S., 2005. Basic science of articular cartilage and osteoarthritis. Clin. Sports Med. 24, 1–12. Robbins, S., Birmingham, T., Callaghan, J., Jones, G., Chesworth, B., Maly, M., 2011a. Association of pain with frequency and magnitude of knee loading in knee osteoarthritis. Arthritis Care Res. 63, 991–997.
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
M.R. Maly et al. / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ Robbins, S., Birmingham, T., Jones, G., Callaghan, J., Maly, M., 2009. Developing an estimate of daily cumulative loading for the knee: examining test–retest reliability. Gait Posture 30, 497–501. Robbins, S., Birmingham, T., Maly, M., Chesworth, B., Giffin, J., 2011b. Comparative diagnostic accuracy of knee adduction moments in knee osteoarthritis: a case for not normalizing to body size. J. Biomech. 44, 968–971. Robertson, D., Dowling, J., 2003. Design and responses of Butterworth and critically damped digital filters. J. Electromyogr. Kinesiol. 13, 569–573. Schache, A., Baker, R., 2007. On the expression of joint moments during gait. Gait Posture 25, 440–452. Schneider, E., Nevitt, M., McCulloch, C., Cicuttini, F., Duryea, J., Eckstein, F., TamezPena, J., 2012. Equivalence and precision of knee cartilage morphometry between different segmentation teams, cartilage regions, and MR acquisitions. Osteoarthr. Cartil. 20, 869–879. Shive, M., Restrepo, A., Totterman, S., Tamez-Pena, J., Schreyer, E., Steinwachs, M., Stanish, W., 2014. Quantitative 3D MRI reveals limited intra-lesional bony overgrowth at 1 year after microfracture-based cartilage repair. Osteoarthr. Cartil. 22, 800–804. Song, J., Semanik, P., Sharma, L., Chang, R., Hochberg, M., Mysiw, W., Bathon, J., Eaton, C., Jackson, R., Kwoh, C., Nevitt, M., Dunlop, D., 2010. Assessing physical activity in persons with knee osteoarthritis using accelerometers: Data from the Osteoarthritis Initiative. Arthritis Care Res. 62, 1724–1732. Stehling, C., Lane, N., Nevitt, M., Lynch, J., McCulloch, C., Link, T., 2010. Subjects with physical activity levels have more severe focal knee lesions diagnosed with 3T MRI: analysis of a non-symptomatic cohort of the osteoarthritis initiative. Osteoarthr. Cartil. 18, 776–786. Tamez-Pena, J., Farber, J., Gonzalez, P., Schreyer, E., Schneider, E., Totterman, S., 2012. Unsupervised segmentation and quantification of anatomical knee
7
features: data from the osteoarthritis initiative. IEEE Trans. Biomed. Eng. 59, 1177–1186. Thorp, L., Sumner, R., Block, J., Moisio, K., Shott, S., Wimmer, M., 2006. Knee joint loading differs in individuals with mild compared with moderate medial knee osteoarthritis. Arthritis Rheum. 54, 3840–3842. Thorp, L., Sumner, R., Wimmer, M., Block, J., 2007. Relationship between pain and medial knee joint loading in mild radiographic knee osteoarthritis. Arthritis Care Res. 57, 1254–1260. Trost, S., McIver, K., Pate, R., 2005. Conducting accelerometer-based activity assessments in field-based research. Med. Sci. Sports Exerc. 37, S531–543. Urquhart, D., Soufan, C., Teichtahl, A., Wluka, A., Hanna, F., Cicuttini, F., 2008. Factors that may mediate the relationship between physical activity and the risk for developing knee osteoarthritis. Arthritis Res. Ther. 10, 203. Vanwanseele, B., Eckstein, F., Smith, R., Lange, A., Foroughi, N., Baker, M., Shnier, R., Fiatarone Singh, M., 2010. The relationship between knee adduction moment and cartilage and meniscus morphology in women with osteoarthritis. Osteoarthr. Cartil. 18, 894–901. Vignon, E., VAlat, J., Rossignol, M., Avouac, B., Rozenberg, S., Thoumie, P., Avouac, J., Nordin, M., Hilliquin, P., 2006. Osteoarthritis of the knee and hip and activity: a systematic international review and synthesis (OASIS). Jt. Bone Spine 73, 442–455. Winter, D., 1990. Biomechanics and Motor Control of Human Movement. WileyInterscience Publication, New York. Wu, G., Cavanagh, P., 1995. ISB recommendations for standardization in the reporting of kinematic data. J. Biomech. 28, 1257–1261.
Please cite this article as: Maly, M.R., et al., Knee adduction moment relates to medial femoral and tibial cartilage morphology in clinical knee osteoarthritis. Journal of Biomechanics (2015), http://dx.doi.org/10.1016/j.jbiomech.2015.04.039i
