In situ measurement of articular cartilage deformation in intact femoropatellar joints under static loading

The deformational behavior of articular cartilage has been investigated in confined and unconfined compression experiments and indentation tests, but to date there exist no reliable data on the in situ deformation of the cartilage during static loading. The objective of the current study was to perform a systematic study into cartilage compression of intact human femoro-patellar joints under short- and long-term static loading with MR imaging. A non-metallic pneumatic pressure device was used to apply loads of 150% body weight to six joints within the extremity coil of an MRI scanner. The cartilage was delineated during the compression experiment with previously validated 2D and 3D fat-suppressed gradient echo sequences. We observed a mean (maximal) in situ deformation of 44% (57%) in patellar cartilage after 32 h of loading (mean contact pressure 3.6 MPa), the femoral cartilage showing a smaller amount of deformation than the patella. However, only around 7% of the final deformation (3% absolute deformation) occurred during the first minute of loading. A 43% fluid loss from the interstitial patellar matrix was recorded, the initial fluid flux being 0.217 +/- 0.083 microm/s, and a high inter-individual variability of the deformational behavior (coefficients of variation 11-38%). In conjunction with finite-element analyses, these data may be used to compute the load partitioning between the solid matrix and fluid phase, and to elucidate the etiologic factors relevant in mechanically induced osteoarthritis. They can also provide direct estimates of the mechanical strain to be encountered by cartilage transplants.     

Muscle recruitment strategies can reduce joint loading during level walking

Joint inflammation, with consequent cartilage damage and pain, typically reduces functionality and affects activities of daily life in a variety of musculoskeletal diseases. Since mechanical loading is an important determinant of the disease process, a possible conservative treatment is the unloading of joints. In principle, a neuromuscular rehabilitation program aimed to promote alternative muscle recruitments could reduce the loads on the lower-limb joints during walking. The extent of joint load reduction one could expect from this approach remains unknown. Furthermore, assuming significant reductions of the load on the affected joint can be achieved, it is unclear whether, and to what extent, the other joints will be overloaded. Using subject-specific musculoskeletal models of four different participants, we computed the muscle recruitment strategies that minimised the hip, knee and ankle contact force, and predicted the contact forces such strategies induced at the other joints. Significant reductions of the peak force and impulse at the knee and hip were obtained, while only a minimal effect was found at the ankle joint. Adversely, the peak force and the impulse in non-targeted joints increased when aiming to minimize the load in an adjacent joint. These results confirm the potential of alternative muscle recruitment strategies to reduce the loading at the knee and the hip, but not at the ankle. Therefore, neuromuscular rehabilitation can be targeted to reduce the loading at affected joints but must be considered carefully in patients with multiple joints affected due to the potential adverse effects in non-targeted joints.     

Comparison of nonlinear mechanical properties of bovine articular cartilage and meniscus

Nonlinear, linear and failure properties of articular cartilage and meniscus in opposing contact surfaces are poorly known in tension. Relationships between the tensile properties of articular cartilage and meniscus in contact with each other within knee joints are also not known. In the present study, rectangular samples were prepared from the superficial lateral femoral condyle cartilage and lateral meniscus of bovine knee joints. Tensile tests were carried out with a loading rate of 5 mm/min until the tissue rupture. Nonlinear properties of the toe region, linear properties in larger strains, and failure properties of both tissues were analysed. The strain-dependent tensile modulus of the toe region, Young's modulus of the linear region, ultimate tensile stress and toughness were on average 98.2, 8.3, 4.0 and 1.9 times greater (p<0.05) for meniscus than for articular cartilage. In contrast, the toe region strain, yield strain and failure strain were on average 9.4, 3.1 and 2.3 times greater (p<0.05) for cartilage than for meniscus. There was a significant negative correlation between the strain-dependent tensile moduli of meniscus and articular cartilage samples within the same joints (r=−0.690, p=0.014). In conclusion, the meniscus possesses higher nonlinear and linear elastic stiffness and energy absorption capability before rupture than contacting articular cartilage, while cartilage has longer nonlinear region and can withstand greater strains before failure. These findings point out different load carrying demands that both articular cartilage and meniscus have to fulfil during normal physiological loading activities of knee joints.     

In Vivo Dynamic Deformation of Articular Cartilage in Intact Joints Loaded by Controlled Muscular Contractions

When synovial joints are loaded, the articular cartilage and the cells residing in it deform. Cartilage deformation has been related to structural tissue damage, and cell deformation has been associated with cell signalling and corresponding anabolic and catabolic responses. Despite the acknowledged importance of cartilage and cell deformation, there are no dynamic data on these measures from joints of live animals using muscular load application. Research in this area has typically been done using confined and unconfined loading configurations and indentation testing. These loading conditions can be well controlled and allow for accurate measurements of cartilage and cell deformations, but they have little to do with the contact mechanics occurring in a joint where non-congruent cartilage surfaces with different material and functional properties are pressed against each other by muscular forces. The aim of this study was to measure in vivo, real time articular cartilage deformations for precisely controlled static and dynamic muscular loading conditions in the knees of mice. Fifty and 80% of the maximal knee extensor muscular force (equivalent to approximately 0.4N and 0.6N) produced average peak articular cartilage strains of 10.5±1.0% and 18.3±1.3% (Mean ± SD), respectively, during 8s contractions. A sequence of 15 repeat, isometric muscular contractions (0.5s on, 3.5s off) of 50% and 80% of maximal muscular force produced cartilage strains of 3.0±1.1% and 9.6±1.5% (Mean ± SD) on the femoral condyles of the mouse knee. Cartilage thickness recovery following mechanical compression was highly viscoelastic and took almost 50s following force removal in the static tests.     

Contact models of repaired articular surfaces: influence of loading conditions and the superficial tangential zone

The superficial tangential zone (STZ) plays a significant role in normal articular cartilage’s ability to support loads and retain fluids. To date, tissue engineering efforts have not replicated normal STZ function in cartilage repairs. This finite element study examined the STZ’s role in normal and repaired articular surfaces under different contact conditions. Contact area and pressure distributions were allowed to change with time, tension-compression nonlinearity modeled collagen behavior in the STZ, and nonlinear geometry was incorporated to accommodate finite deformation. Responses to loading via impermeable and permeable rigid surfaces were compared to loading via normal cartilage, a more physiologic condition, anticipating the two rigid loading surfaces would bracket that of normal. For models loaded by normal cartilage, an STZ placed over the inferior repair region reduced the short-term axial compression of the articular surface by 15%, when compared to a repair without an STZ. Covering the repair with a normal STZ shifted the flow patterns and strain levels back toward that of normal cartilage. Additionally, reductions in von Mises stress (21%) and an increase in fluid pressure (13%) occurred in repair tissue under the STZ. This continues to show that STZ properties of sufficient quality are likely critical for the survival of transplanted constructs in vivo. However, response to loading via normal cartilage did not always fall within ranges predicted by the rigid surfaces. Use of more physiologic contact models is recommended for more accurate investigations into properties critical to the success of repair tissues.     

In situ chondrocyte deformation with physiological compression of the feline patellofemoral joint

The mechanical environment is an important factor affecting the maintenance and adaptation of articular cartilage, and thus the function of the joint and the progression of joint degeneration. Recent evidence suggests that cartilage deformation caused by mechanical loading is directly associated with deformation and volume changes of chondrocytes. Furthermore, in vitro experiments have shown that these changes in the mechanical states of chondrocytes correlate with a change in the biosynthetic activity of cartilage cells. The purpose of this study was to apply our knowledge of contact forces within the feline patellofemoral joint to quantify chondrocyte deformation in situ under loads of physiological magnitude. A uniform, static load of physiological magnitude was applied to healthy articular cartilage still fully intact and attached to its native bone. The compressed cartilage was then chemically fixed to enable the evaluation of cartilage strain, chondrocyte deformation and chondrocyte volumetric fraction. Patella and femoral groove articular cartilages differ in thickness, chondrocyte aspect ratio, and chondrocyte volumetric fraction in both magnitude and depth distribution. Furthermore, when subjected to the same compressive loads, changes to all of these parameters differ in magnitude and depth distribution between patellar and femoral groove articular cartilage. This evidence suggests that significant chondrocyte deformation likely occurs during in vivo joint loading, and may influence chondrocyte biosynthetic activity. Furthermore, we hypothesise that the contrasts between patella and femoral groove cartilages may explain, in part, the site-specific progression of osteoarthritis in the patellofemoral joint of the feline anterior cruciate ligament transected knee.     

Contact stress transients during functional loading of ankle stepoff incongruities

Cartilage deformation demonstrates viscoelastic behavior due to its unique structure. However, nearly all contact studies investigating incongruity-associated changes in cartilage surface stresses have been static tests. These tests have consistently measured only modest increases in contact stresses, even with large incongruities. In this study, an experimental approach measuring real-time contact stresses in human cadaveric ankles during quasi-physiologic motion and loading was used to determine how stepoff incongruities of the distal tibia affected contact stresses and contact stress gradients. Peak instantaneous contact stresses, in ankles with stepoffs between 1.0 and 4.0mm of the anterolateral articular surface, increased by between 2.3 x and 3.0 x compared to the corresponding intact ankle values. Peak instantaneous contact stress gradients in stepoff configurations increased by between 1.9 x and 2.6 x the corresponding intact configuration values. Anatomic reduction of the displaced fragment restored intact contact stresses and contact stress gradients. Intact and anatomic configurations demonstrated a heterogeneous population of low-magnitude, randomly oriented contact stress gradient vectors in contrast to high-magnitude, preferentially oriented gradients in stepoff configurations. Peak instantaneous contact stresses may be important pathomechanical determinants of post-traumatic arthritis. Abnormal contact stress gradients could cause regional pathological disturbances in cartilage stress and interstitial fluid distribution. Measuring contact stresses and contact stress gradients during motion allowed potential incongruity-associated pathologic changes in loading that occur over the complete motion cycle to be investigated.     

Comparison of lower limb kinetics,kinematics and muscle activation during drop jumping under shod and barefoot conditions

This study was to investigate the acute effects of wearing shoes on lower limb kinetics, kinematics and muscle activation during a drop jump. Eighteen healthy men performed a drop jump under barefoot and shod conditions. Vertical ground reaction force (GRF) was measured on a force plate during the contact phase of a drop jump, and GRF valuables were calculated for each condition. The angles of the knee and ankle joints, and the foot strike angle (the angle between the plantar surface of the foot and the ground during ground contact) as well as the electromyography of 7 muscles were measured. The shod condition showed a significant larger first peak GRF, longer time to first peak GRF from the initial ground contact and lower initial loading rate than the barefoot condition. The shod condition showed a significant larger ankle joint angle at initial ground contact, smaller knee joint angle between the second peak GRF and take-off as well as smaller foot strike angle at both initial ground contact and take-off than the barefoot condition. There were significant correlations between relative differences in ankle joint at the initial ground contact and relative differences in the initial loading rate. The muscle activity of all muscles during foot ground contact did not differ between two conditions; however, in the shod condition, muscle activation of 150 ms before foot ground contact was significantly higher in the rectus femoris, whereas it was lower in the biceps femoris and tibialis anterior muscles than the barefoot condition. These results indicate that wearing shoes alternates the GRF variables at initial ground contact, joint kinematics at the ground contact and muscle activation before foot ground contact during a drop jump, suggesting that the effects of wearing shoes on drop jump training differ from being barefoot.     

Contact area and pressure distribution in the feline patellofemoral joint under physiologically meaningful loading conditions.

The purpose of this study was to determine contact area and mean and peak pressures in the healthy feline patellofemoral joint over the complete range of possible applied force. Furthermore, we wanted to improve upon the repeatability of previous measurements while maximizing the physiological relevance of the results obtained. The patellae and femora were secured in a loading frame approximating an in situ loading configuration. Low- and medium-grade Fuji film was used to assess patellofemoral contact area and pressure distribution, respectively. Constant force was applied to the patellofemoral joints for 2s (short duration trials) or 5min (long duration trials). For the short duration trials, contact area was shown to increase logarithmically with the force applied. In contrast, mean and peak pressures increased linearly with force. Furthermore, the rate of increase of peak pressure with force was approximately three times greater than that of mean pressure. For the long duration trials, contact area increased up to 33% compared to the short duration trials. This effect could no longer be detected with our approach after an unloading period of 5-10s. Increasing contact area is one mechanism that the feline patellofemoral joint may use to regulate the pressures experienced by the cartilage as the force applied to the joint increases. The attenuation of external forces inside a joint is achieved by the specific geometry of the articulating surfaces and the viscoelastic properties of the articular cartilage. It likely represents a natural protection of joints to high external load magnitudes.     

The transmission of load through the human hip joint

This paper describes the results of loading experiments carried out on human hip joints. The unloaded surfaces of the femoral head and the acetabulum are slightly incongruous. The location and magnitude of the contact areas between the surfaces therefore depend on the magnitude and direction of the applied load. The contact areas were determined experimentally for a variety of loads typical of normal walking. Two distinct contact areas were found on the anterior and posterior aspects of the acetabulum at light loads, gradually merging with increasing load until, at a certain transition load, the dome of the acetabulum comes into contact and contact is then complete. The value of the transition load depends on the rate of loading, due to creep of the cartilage, and was found to vary from 50 per cent of body weight in young specimens to 25 per cent of body weight for elderly specimens for rates of loading typical of normal walking. Thus, the dome of the acetabulum is out of contact for a substantial portion of the swing phase of normal walking.

The analysis of a much simplified model of the hip joint is presented. The dependence of contact area on load is demonstrated, but also a method of determining the transition load for complete contact from the load/deflection relation for the hip is suggested. The values of the transition load quoted above were obtained by this method. The analysis further indicates that the distribution of pressure between the articular surfaces depends critically on the distribution of cartilage thickness throughout the joint. It is suggested that the distribution of cartilage thickness is such as to lead to a state of uniform pressure at the upper end of the physiological load range. Some experimental evidence is presented in support of this suggestion.

It is concluded that the function of joint incongruity is to allow the articular surfaces to come out of contact at light loads so that the cartilage may be exposed to synovial fluid for the purposes of nutrition and lubrication. At large loads, the distribution of cartilage thickness ensures that a state of hydrostatic pressure is achieved in order that cartilage, with a large fluid content, may transmit large pressures without flow and consequent loss of its integrity.     

Deformation and recovery of cartilage in the intact hip under physiological loads using 7 T MRI

Accurate measurement of cartilage deformation in loaded cadaver hip joints could be a valuable tool to answer clinically relevant research questions. MRI is a promising tool, but its use requires an understanding of cartilage deformation and recovery properties in the intact hip. Our objective was to answer the following questions: (1) How long does it take for hip cartilage to reach a deformed steady-state thickness distribution under simulated physiological load, and how much does the cartilage deform? (2) How long does it take for hip cartilage to return to the original cartilage thickness distribution once the load is removed?MethodsFive human hip specimens were axially loaded to 1980 N in a 7 T MR scanner and scanned every 15 min throughout loading. One specimen was scanned every hour throughout recovery from load. One repeatability specimen was loaded and scanned every day for 4 days. Hip cartilage was segmented as a single unit and thickness was measured radially.ResultsThe hip cartilage reached a steady-state thickness distribution after 225 min of load, and 16.5 h of recovery. Mean strain after 225 min of load was 30.9%. The repeatability specimen showed an average day-to-day change in mean cartilage thickness of 0.10 mm over 4 days of data collection. The amount of deformation (0.96 mm) was far greater than the image resolution (0.11 mm) and error due to repeatability (0.10 mm).ConclusionUsing an ex vivo model, this method has potential for assessing changes in hip cartilage strain due to injury or surgical intervention.     

Two-dimensional strain fields on the cross-section of the bovine humeral head under contact loading

The objective of this study was to provide a detailed experimental assessment of the two-dimensional cartilage strain distribution on the cross-section of immature and mature bovine humeral heads subjected to contact loading at a relatively rapid physiological loading rate. Six immature and six mature humeral head specimens were loaded against glass and strains were measured at the end of a 5s loading ramp on the textured articular cross-section using digital image correlation analysis. The primary findings indicate that elevated tensile and compressive strains occur near the articular surface, around the center of the contact region. Few qualitative or quantitative differences were observed between mature and immature joints. Under an average contact stress of approximately 1.7 MPa, the peak compressive strains averaged -0.131+/-0.048, which was significantly less than the relative change in cartilage thickness, -0.104+/-0.032 (p<0.05). The peak tensile strains were significantly smaller in magnitude, at 0.0325+/-0.013. These experimental findings differ from a previous finite element analysis of articular contact, which predicted peak strains at the cartilage-bone interface even when accounting for the porous-hydrated nature of the tissue, its depth-dependent inhomogeneity, and the disparity between its tensile and compressive properties. These experimental results yield new insights into the local mechanical environment of the tissue and cells, and suggest that further refinements are needed in the modeling of contacting articular layers.     

In vivo cartilage contact strains in patients with lateral ankle instability

Damage to the anterior talofibular ligament (ATFL) and cacaneofibular ligament (CFL) during an ankle sprain may be linked to the development of osteoarthritis. Although altered tibiotalar kinematics have been demonstrated, the effects of lateral ankle instability (LAI) on in vivo cartilage strains have not been described. We hypothesized that peak cartilage strains increase, and the location is shifted in patients with ATFL injuries. We used 3-D MRI models and biplanar fluoroscopy to evaluate in vivo cartilage contact strains in seven patients with unilateral LAI. Subjects had chronic unilateral ATFL injury or combined ATFL and CFL injury, and were evaluated with increasing load while stepping onto a force plate. Peak cartilage strain and the location of the peak strain were measured using the contralateral normal ankle as a control. Ankles with LAI demonstrated significantly increased peak strain when compared with ATFL-intact controls. For example, at 100% body weight, peak strain was 29±8% on the injured side compared to 21±5% on the intact side. The position of peak strain on the injured ankle also showed significant anterior translation and medial translation. At 100% body weight, the location of peak strain in the injured ankle translated anteriorly by 15.5±7.1 mm and medially by 12.9±4.3 mm relative to the intact ankle. These changes correspond to the region of clinically observed osteoarthritis. Chronic LAI, therefore, may contribute to the development of tibiotalar cartilage degeneration due to altered cartilage strains.     

Load sharing between solid and fluid phases in articular cartilage: II--Comparison of experimental results and u-p finite element predictions.

Experimental measurements in conjunction with theoretical predictions were used to determine the extent of load supported by the fluid phase of cartilage at the articular surface. The u-p finite element model was used to simulate the loading of six separate porcine knee joints and to predict surface deformations of the cartilage layer on the lateral femoral condyle. Representative geometry for the condyle, contact pressures, and intrinsic material properties of the cartilage layer were supplied from experimental measures (see Part I). The u-p finite element predictions for surface deformations of the cartilage layer were obtained for several load partitioning states between the solid and fluid phases of cartilage at the articular surface. These were then compared to actual surface deformations obtained experimentally. It appeared from the comparison that approximately 75 percent of the applied load was borne by the fluid phase at the articular surface under this loading regime. This was qualitatively in agreement with the hypothesis that an applied load to articular joints is partitioned at the surface to the two phases according to the surface area ratios of the solid and fluid phases. It appeared that the solid phase was shielded from the total applied stress on the articular surface by the fluid and could be a reason for the excellent durability of the tissue under the demanding conditions in a diarthrodial joint.     

Changes in articular cartilage mechanics with meniscectomy: A novel image-based modeling approach and comparison to patterns of OA

Meniscectomy is a significant risk factor for osteoarthritis, involving altered cell synthesis, central fibrillation, and peripheral osteophyte formation. Though changes in articular cartilage contact pressure are known, changes in tissue-level mechanical parameters within articular cartilage are not well understood. Recent imaging research has revealed the effects of meniscectomy on the time-dependent deformation of physiologically loaded articular cartilage. To determine tissue-level cartilage mechanics that underlie observed deformation, a novel finite element modeling approach using imaging data and a contacting indenter boundary condition was developed. The indenter method reproduces observed articular surface deformation and avoids assumptions about tangential stretching. Comparison of results from an indenter model with a traditional femur-tibia model verified the method, giving errors in displacement, solid and fluid stress, and strain below 1% (RMS) and 7% (max.) of the absolute maximum of the parameters of interest. Indenter finite element models using real joint image data showed increased fluid pressure, fluid exudation, loss of fluid load support, and increased tensile strains centrally on the tibial condyle after meniscectomy-patterns corresponding to clinical observations of cartilage matrix damage and fibrillation. Peripherally there was decreased consolidation, which corresponds to reduced contact and fluid pressure in this analysis. Clinically, these areas have exhibited advance of the subchondral growth front, biological destruction of the cartilage matrix, cartilage thinning, and eventual replacement of the cartilage via endochondral ossification. Characterizing the changes in cartilage mechanics with meniscectomy and correspondence with observed tissue-level effects may help elucidate the etiology of joint-level degradation seen in osteoarthritis.     

Load sharing between solid and fluid phases in articular cartilage: I--Experimental determination of in situ mechanical conditions in a porcine knee.

The in situ mechanical conditions of cartilage in the articulated knee were quantified during joint loading. Six porcine knees were subjected to a 445 N compressive load while cartilage deformations and contact pressures were measured. From roentgenograms, cartilage thickness before and during loading allowed the calculation of tissue deformation on the lateral femoral condyle at different times during the loading process. Contact pressures on the articular surface were measured with miniature fiber-optic pressure transducers. Results showed that the medial side of the lateral femoral condyle had higher contact pressures, as well as deformations. To begin to correlate the pressures and resulting deformations, the intrinsic material properties of the cartilage on the lateral condyle were obtained from indentation tests. Data from four normal control specimens indicated that the aggregate modulus of the medial side was significantly higher than in other areas of the condyle. These experimental measures of the in situ mechanical conditions of articular cartilage can be combined with theoretical modeling to obtain valuable information about the relative contributions of the solid and fluid phases to supporting the applied load on the cartilage surface (see Part II).     

The mechanical environment of chondrocytes in articular cartilage

There is a growing literature concerning chondrocyte responses to mechanical loading, but relatively little is known about the mechanical environment these cells experience in a living joint. Calculations indicate that high forces are applied to limb joints whenever the joints are flexed, because flexion can cause body weight to act on long lever arms compared to the joint centre, whereas the muscles which extend the joint act on much shorter lever arms. As a result, joint reaction forces (which compress the cartilage) can rise to 3-6 times body weight during activities such as stair climbing. Articular cartilage tends to spread this load evenly over the joint surface, but is too thin to do this well, and compressive stresses can rise to 10-20 MPa. Within cartilage, matrix stresses vary locally, possibly as a result of variation in composition or undulations in the subchondral bone, and further modifications of stress occur within each chondron. Articular cartilage is a fibrous solid and cells within it are deformed by mechanical loading rather than subjected to a hydrostatic pressure. The mechanical environment of chondrocytes can best be reproduced in vitro by direct compression of the articular surface of cartilage which is supported naturally by adjacent cartilage and subchondral bone.     

A note on an asymptotic solution for the contact of two biphasic cartilage layers in a loaded synovial joint at rest.

Retaining the first terms of asymptotic expansions and assuming zero gradient of the contact pressure at the contact edge in the perpendicular direction to the edge, Ateshian et al. (1994, J. Biomechanics 27, 1347-1360; 1992, Adv. Bioeng. ASME 22, 191-194) have presented an asymptotic axially symmetric and plane strain solutions to the contact problems of the two thin biphasic cartilage layers in the synovial joint (with the synovial fluid film in between already depleted), subjected to a sudden normal load. Both the immediate and an early time response to a step loading have been analysed. The present note shows that the contact width thus obtained immediately after the load application differs from the numerical values for a dry frictionless contact of incompressible single-phase elastic layers available in the literature. The difference increases with the decreasing contact width-to-layer thickness ratio. It is proposed to improve the above solutions by releasing zero pressure gradient condition at the contact edge and calculating the instantaneous contact widths from the equations proposed for the corresponding case by Matthewson (1981, J. Mech. Phys. Solids 29, 89-113) and Meijers (1967, Appl. Sci. Res. 18, 353-383) that approximate better the numerical values. The instantaneous rate of change in the contact width is then obtained for a step load varying in time.     

Analysis of the Relationship between Peak Stress and Proteoglycan Loss following Injurious Compression of Human Post-mortem Knee and Ankle Cartilage

While traumatic joint injuries are known to increase the risk of osteoarthritis (OA), the mechanism is not known. Models for injurious compression of cartilage may identify predictors of injury that suggest a clinical mechanism. We investigated the relationship between peak stress during compression and glycosaminoglycan (GAG) loss after injury for knee and ankle cartilages. Human cartilage explant disks were harvested post-mortem from the knee and ankle of three organ donors with no history of OA and subjected to injurious compression to 65% strain in uniaxial unconfined compression at 2 mm/s (400%/s). The GAG content of the conditioned medium was measured 3 days after injury. After injury of knee cartilage disks, damage was visible in 18 of 39 disks (36%). Three days after injury, the increase in GAG loss to the medium (GAG loss from injured disks minus GAG loss from location-matched uncompressed controls) was 1.5±0.3 μg/disk (mean ± SEM). With final strain and compression velocity held constant, we observed that increasing peak stress during injury was associated with less GAG loss after injury (P<0.001). In contrast, ankle cartilage appeared damaged after injury in only 1 of 16 disks (6%), there was no increase in GAG loss (0.0±0.3 μg/disk), and no relationship between peak stress and increase in GAG loss was detected (P=0.51). By itself, increasing peak stress did not appear to be an important cause of GAG loss from human cartilage in our injurious compression model. However, we observed further evidence for differences in the response of knee and ankle cartilages to injury.