Knee biomechanics during impact landing understanding injury mechanisms and developing prevention strategies 3

The exponential increase in peak GRF and the relatively slower increase in knee flexion angles, angular velocities and joint power may synergistically lead to an exacerbated lower extrem

CHAPTER 6 DISCUSSIONS

~ It is more noble by silence to avoid an injury than by argument to overcome it ~

Francis Beaumont

6 DISCUSSIONS

6.1 Understanding Biomechanics of Landing

In view of previous landing studies, the relationships of landing height with GRF, knee flexion angles, angular velocities and joint powers are not well understood and there is an existing limitation in obtaining these data at very large landing height

in a controlled laboratory setting The regression work in Stage A aimed to investigate these relationships through a series of landing tasks from a range of landing heights (0.15-1.05m) which were commonly tested; these relationships may be employed to predict these parameters at large landing heights

The range of peak GRF obtained in this stage was slightly lower compared to previous studies (Mizrahi and Susak, 1982; Dufek and Bates, 1990; McNair and Marshall, 1994; McNitt-Gray, 1993; Seegmiller and McCaw, 2003; Kernozek et al., 2005; Pappas et al., 2007), though the peak GRF data obtained among all these studies were quite varied This may be explained by the type of landing style adopted

by the different subjects in these studies DeVita and Skelly (1992) stated that a soft landing style produces a mitigated GRF compared to stiff landing Moreover, it was previously reported that gymnasts who tend to land with minimal knee flexion incurred higher GRF than recreational athletes (Seegmiller and McCaw, 2003) In this study, it was notable to observe a marked exponential relationship between peak resultant GRF and landing height

The presence of larger peak impact force indicated weaker shock-absorbing capacity, which may increase the susceptibility to lower-extremity overuse injuries (Bus, 2003; Hargrave et al., 2003) The results established that the peak GRF incurred during landing is highly dependent on landing height and the exponentially-elevated

peak GRF during landing from a great height can potentially heighten the risk of developing lower extremity injuries

The knee flexion angles at initial contact and at peak GRF demonstrated inverse-exponential relationship with landing height The gradual increase in knee flexion angles at both events with landing height up to 0.75m may represent the inherent mechanism of the musculoskeletal system to increase knee flexion so as to provide enhanced shock absorption capacity against the elevated GRF (Hargrave et al., 2003; Coventry et al., 2006) However, at landing heights beyond 0.75m, the knee flexion angles at initial contact and at peak GRF were found to increase at a slower rate The reduced rate of increase observed in these knee flexion angles may contribute to diminished shock absorption in response to exponentially increasing peak GRF

DeVita and Skelly (1992) reported that a soft landing style leads to an attenuated GRF relative to stiff landing Thus, the presence of lower knee flexion during stiff landing can lead to a diminished shock absorption capacity, whereby the articular cartilage will sustain large compressive impact loads that can afflict cartilage lesions (Lafortune et al., 1996; Hargrave et al., 2003) Yu et al (2006) previously demonstrated that the knee flexion angular velocity has a negative correlation with peak GRF during landing, which implied that active knee flexion, in the form of higher rate of increase in knee flexion with time, is an important player in impact force attenuation Furthermore, the presence of a negative knee joint power revealed eccentric work done on knee extensors to dissipate impact energy Zhang et al (2000) reported that negative knee joint power increased with landing height, which suggested that the knee extensors were vital contributors to energy dissipation Additionally, in the current study, an inverse-exponential relationship was noted

between knee flexion angular velocity and landing height, and a simple linear relationship between knee joint power and landing height Altogether, these findings suggested that the energy dissipation capacity of the knee joint increased at a relatively slower rate at higher landing heights despite the exponentially-increasing GRF parameters This ‘misbalance’ between energy dissipation capacity and GRF parameters at great landing heights is likely to aggravate lower extremity injury risk

One key limitation of this study is the progressive conduct of the landing tasks with regards to the landing height Potential injury risks may exist in randomized landing trials; for instance, if the subjects were to perform the landing trials with greater height (0.90-1.05 m) initially, there is a higher tendency for injury compared

to having the landing trials with lower height (0.15-0.30 m) conducted first Though the findings of this study may not be completely reflective of the actual landing condition wherein an athlete lands from a large height in a single task, this progressive landing protocol was beneficial as it helps the subjects to attune to the landing style that best facilitates them in shock attenuation

Another limitation is perhaps the potential inter-subject differences in landing style, which may influence the regression relationships obtained for each subject between the dependent and independent variables Although the subjects were not specificially instructed to follow a standardized landing style, a qualitative examination of their landing motions revealed mostly ‘soft’ landing styles which may explain the lower peak GRF obtained compared to previous studies Since the current study was more concerned with investigating the effect of landing height on GRF, knee flexion angles, angular velocities and joint powers for recreational athletes who have different landing experiences, the subjects were permitted to execute their natural landing styles Setting a standardized landing style for all subjects may

introduce confounding variables such as the individual ability in learning the landing style; a standardized landing style would be more relevant for studies on specific groups of athletes, such as gymnasts and volleyball players, who would possess similar levels of learning ability and landing experience

It is important to note that the study was constrained by a small sample size, hence the results may not be representative of the general population However, the post-hoc power analysis revealed that there was generally a sufficiently high power to detect the observed regression relationships in the data obtained in this study While it cannot be concluded the general population would follow these regression relationships during landing, the study was able to show that the relationships have a strong fit in terms of R2 and p values for the 5 subjects tested in this regression study

In addition, there are certain factors in a physiological landing task specific to a sports-/military-related setting, like upper body motion, friction/gradient of landing surface, type of shoe/boots worn, additional carried weights and post-landing maneuvers, which are not examined in the current study These factors may usually cause unexpected effects on the landing biomechanics of the athlete and elevate injury risk (Dufek and Bates, 1991) Since the intention of this study was to investigate the regression relationships of landing height with GRF, knee flexion angles, angular velocities and joint powers, it is therefore necessary to conduct the present study in a controlled laboratory environment so that the subjects knew exactly what to expect and can perform the same sets of landing tasks with minimal injury risk

We further acknowledged that one limitation of this study was, perhaps, the usage of male subjects This project was performed in collaboration with the Defence Medical and Environmental Research Institute, which necessitated the testing of male subjects as the local males (at the age of 18) are subjected to military conscription and

form a dominant population of our armed forces Hence, there is a need to identify the biomechanical risk factors for lower extremity injuries in our male soldiers, who are regularly required to perform landing tasks from obstacles of various heights during military training or exercises

The results of this study collectively established that the peak resultant GRF possessed a strong exponential regression relationship with landing height, while knee flexion angles (at initial contact and peak GRF) and peak knee flexion angular velocities followed an inverse-exponential relationship with landing height, and peak knee joint power adopted a simple linear regression relationship with landing height The parameters analyzed in this study are highly dependent on landing height The exponential increase in peak GRF and the relatively slower increase in knee flexion angles, angular velocities and joint power may synergistically lead to an exacerbated lower extremity injury risk at large landing heights

The prior regression study revealed how the landing height can directly influence the magnitude of peak GRF, knee flexion angles, angular velocities and joint power; however, based on previous landing studies, there is still a lack of understanding of how the knee joint will respond in terms of the kinematics and energetics to the combined effects of different landing heights and techniques The next study in Stage A sought to examine the knee joint kinematics and energetics sustained during landing phase, in response to the effects of landing height (0.3m and 0.6m) and technique (single-leg and double-leg landing) The key observations of the study were as follows: 1) heightened peak GRF during single-leg landing and/or at greater landing height, 2) higher knee flexion angles and angular velocities during double-leg landing and/or at greater landing height, and 3) elevated joint power and eccentric work done during double-leg landing and/or at larger landing height

The range of peak GRF obtained in the current study is consistent with several previous studies that examined GRF during single-leg and double-leg landing techniques Pappas et al (2007) observed a higher GRF at 40-deg knee flexion for the dominant limb during single-leg landing than double-leg landing from a 0.4-m height Similarly, in the current study, a greater peak resultant GRF was obtained for single-leg landing than double-leg landing for both 0.3-m and 0.6-m landing heights McNitt-Gray (1993), and Seegmiller and McCaw (2003) reported an increase in peak GRF when the landing height was elevated during double-leg landing The findings added that the aggravation in peak GRF with increased landing height was generally present for both double-leg and single-leg landing, though it was statistically significant only during single-leg landing The greater GRF noted for single-leg landing from both landing heights suggested that the single-leg landing technique may involve a larger lower extremity injury risk, relative to double-leg landing

Double-leg landing generally allows elevated knee flexion angles throughout the landing phase of double-leg landing compared to single-leg landing (Pappas et al., 2007); moreover, Madigan and Pidcoe (2003) reported that the range of motion for knee flexion were greater for double-leg landings The results revealed that this elevation in knee flexion angles was further enhanced when the landing height was increased This observation indicated that the knee joint is able to respond more effectively in terms of kinematics against GRF during double-leg landing, compared

to single-leg landing The response was exhibited in the form of increased knee flexion angles during landing phase in order to promote shock attenuation as the knee joint is known to be partly responsible for the body's ability to absorb shock during ground contact (DeVita and Skelly, 1992; Hargrave et al., 2003)

On the other hand, the presence of a larger peak GRF during single-leg landing is attributed, in part, to a diminished capacity for shock attenuation DeVita and Skelly (1992) reported that a soft landing style produces a mitigated GRF compared to stiff landing Seegmiller and McCaw (2003) further suggested gymnasts tended to land with minimal knee flexion and thus incurred higher GRF than recreational athletes Hence, the presence of lower knee flexion during single-leg landing can lead to a reduced shock absorption capacity, wherein the knee joint will sustain large compressive impact loads (Jeffrey and Aspden, 2006; Lafortune et al., 1996; Yeow et al., 2008) which can potentially inflict cartilage lesions (Coventry et al., 2006; Madigan and Pidcoe, 2003; Hargrave et al., 2003) and also place the ACL at

a higher risk for injury (Lephart et al., 2002; Olsen et al., 2004; Yeow et al., 2008; Boden et al., 2000)

In terms of knee flexion angular velocity, double-leg landing permitted a substantially greater angular velocity compared to single-leg landing; furthermore, it was noted that the angular velocity markedly increased with landing height during double-leg landing These findings demonstrated that the knee joint was able to respond with relatively more immediacy to the landing impact via active knee flexion, during double-leg landing It was also observed that the increment in landing height could also facilitate the escalation of this response Yu et al (2006) have previously demonstrated that the knee flexion angular velocity has a negative correlation with peak GRF during landing, thus indicating that active knee flexion is a key factor in impact force attenuation In this study, the presence of a larger knee flexion angular velocity and lower peak GRF during double-leg landing clearly illustrated enhanced shock attenuation, relative to single-leg landing Furthermore, the increased knee

flexion angular velocity is necessary to cope with the elevated GRF that arises from the increment in landing height

Similar to knee flexion angular velocity, a greater negative knee joint power was found during double-leg landing than single-leg landing; the negative joint power also increased significantly with landing height during double-leg landing DeVita and Skelly (1992) reported that the presence of a negative knee joint power indicated eccentric work done on knee extensors to dissipate impact energy Zhang et al (2000) further illustrated an increase in negative knee joint power with landing height, which suggested that the knee extensors were key contributors to energy dissipation during landing The results showed that the increment in landing height increased the negative joint power during double-leg landing, which implied elevated eccentric work done on the knee extensors

The findings of this study also demonstrated that the eccentric work done was elevated by 90.2% when the landing height increased from 0.3-m to 0.6-m during double-leg landing, but only a smaller increase of 59.9% was noted for single-leg landing In addition, at the 0.3-m landing height, the eccentric work done decreased

by 43.2% during single-leg landing, relative to double-leg landing; the reduction was further aggravated to 52.3% for the 0.6-m landing height The findings suggest that the knee joint was able to respond with greater energy dissipation during double-leg landing than single-leg landing, at both tested landing heights, and this may be attributed to the reduced knee joint kinematics in the latter landing technique

A main limitation of this study is that the present study was conducted in a controlled laboratory environment where the subjects were briefed on the requirements of the trials and knew what to expect Although this study allows reasonable comparison of GRF, knee joint kinematics and energetics between landing

heights and between landing techniques as all subjects performed the same set of tasks, the results may not completely reflect the physiological landing maneuver during sports- or military-related landing activities A physiological landing maneuver can involve unexpected factors, such as friction of landing surface, type of shoe/boots, upper body motion and additional weights carried, that can render the landing more unpredictable and risky than a simple landing technique in a controlled laboratory It should be acknowledged that the motion-capture system has an innate degree of error

in kinematics measurement during rapid human motion Though every effort was made to ensure minimal marker motion, it is still possible that the limitations of the instrumentation may affect the results to some extent

Additionally, the unilateral development of certain overuse and acute lower extremity injuries suggests that the lower extremity function may not be completely bilaterally symmetrical Schot et al (1994) reported the presence of bilateral asymmetry among left and right side vertical GRFs and lower extremity joint moments during double-leg landing from a 0.6-m height It should be noted that the main aim of this study was not to examine the bilateral differences in double-leg landing, but to investigate the effect of different landing techniques on lower extremity biomechanical parameters We therefore selected biomechanical data from the dominant limb to allow consistent comparison between single-leg and double-leg landing tasks

The results collectively indicated that an increase in landing height and/or the use of the single-leg landing technique aggravated the peak GRF during landing phase Moreover, the knee joint displayed greater knee flexion angles at initial contact and at peak GRF, and maximum knee flexion during double-leg landing, compared to single-leg landing The knee joint further exhibited elevated knee flexion angles at

peak GRF and maximum knee flexion angles during double-leg landing with increased landing height The findings illustrated that the knee joint delivered larger knee flexion angular velocity, negative knee joint power and eccentric work done during double-leg landing, than single-leg landing, at both heights Altogether, the knee joint is able to respond more effectively in terms of kinematics and energetics to

a larger landing impact from an elevated height during double-leg landing, compared

to single-leg landing This allows better shock absorption and thus minimizes the risk

of sustaining lower extremity injuries

From the previous two studies in Stage A, it was found that the knee serves to provide shock absorption for the landing impact and the energy dissipation capacity of the knee joint is strongly associated with landing height and landing technique The next question would be – how much does the knee joint contribute to the total energy dissipation compared to other lower extremity joints during impact landing?

Currently, there is limited understanding on the differences in energy dissipation strategies adopted by the lower extremity joints between single-leg and double-leg landing maneuvers The purpose of this next study was to compare the energy dissipation strategies adopted during double-leg and single-leg landing by examining their differences in sagittal and frontal plane energetics Altogether, the results indicated four major findings: (1) the hip and the knee were the major contributors to energy dissipation in the sagittal plane for double-leg landing, (2) the hip and the ankle were the dominant contributors to energy dissipation in the sagittal plane for single-leg landing, (3) the hip was the main contributor to energy dissipation

in the frontal plane for double-leg landing, and (4) the knee was the chief contributor

to energy dissipation in the frontal plane for single-leg landing

In both double-leg and single-leg landing, it was noted that the hip and knee generally achieved greater joint angles at initial contact and peak joint angles than the ankle It could be explained by the limited anatomical range of motion for the ankle joint Furthermore, Chappel et al (2007), Kernozek et al (2008), McNitt-Gray (1993), and Self and Paine (2001) have previously reported that recreational athletes tended to adopt a landing strategy with relatively higher hip and knee flexion It was also observed that the hip and knee attained larger joint angles at initial contact and peak joint angles during double-leg landing than single-leg landing in the sagittal plane; however, this difference was absent in the frontal plane, which might be due to the lesser range of motions available to the lower extremity joints in the frontal plane relative to the sagittal plane (Kernozek et al., 2005)

Interestingly, though the ankle displayed lower sagittal plane joint angles during double-leg and single-leg landing, its joint angular velocities were still relatively high compared to the hip While active knee flexion has been reported to play a vital role in shock absorption (Yeow et al., 2009b; Yu et al., 2006), the results obtained in this current study indicated that the ankle might play a similar role for both double-leg and single-leg landing It was also found that in the frontal plane, the knee exhibited higher joint angular velocity during single-leg landing than double-leg landing, which further underscored the knee joint as a critical player in shock attenuation in the frontal plane

In terms of joint kinetics, the findings illustrated a larger sagittal plane joint moment at the hip during single-leg landing, compared to double-leg landing; however, the frontal plane joint moment at the hip was greater during double-leg landing, compared to single-leg landing This phenomenon was conversely true for the knee joint; the presence of a high knee abduction moment during single-leg

landing can increase anterior cruciate ligament strain and aggravate rupture risk (Shin

et al., 2009) As the external joint moment is indicative of the counter-moment generated by the joint muscles to stabilize the joint, the findings suggested that the hip abductors and knee extensors might be chiefly responsible for applying the stabilizing forces during double-leg landing (Zhang et al., 2000); but during single-leg landing, the hip extensors and knee adductors were the main joint stabilizers The observations made in this study were consistent with a previous report by Dufek and Bates (1990) which reported greater peak moments of the proximal lower extremity muscles compared with the distal muscles

The data on joint power and eccentric work demonstrated 35.1% for the hip, 35.3% for the knee and 29.7% for the ankle in terms of contribution to total energy dissipation in the sagittal plane during double-leg landing However, it should be noted that the contributions increased to 42.9% for the hip, decreased to 11.4% for the knee and increased to 45.7% for the ankle during single-leg landing For the frontal plane, the contributions to total energy dissipation were 66.7% for the hip, 29.0% for the knee and 4.3% for the ankle during double-leg landing During single-leg landing, the contributions decreased to 36.6% for the hip, increased to 60.7% for the knee and decreased to 2.7% for the ankle during single-leg landing These findings emphasized the roles of the hip and knee as the dominant energy dissipaters in the sagittal plane for double-leg landing, which was in line with previous studies (DeVita and Skelly, 1992; Zhang et al., 2000, Decker et al., 2003) The results further indicated that the hip and ankle were the main energy dissipaters in the sagittal plane during single-leg landing As for the frontal plane, the hip and the knee acted as the key energy dissipaters for double-leg and single-leg landing respectively

One constraint of this study was perhaps the use of natural landing style A standardized landing style would offer more experimental consistency for subjects who are trained in the same sports as the subjects would likely share similar levels of landing experience However, in this study, the subjects are recreational athletes and play different sports, therefore they may have different abilities in learning the standardized landing style Moreover, the use of natural landing style benefited us in investigating how the subjects would naturally coordinate their lower extremity joints

in order to achieve an appropriate strategy for absorbing the landing impact

Finally, it was demonstrated in this study that in the sagittal plane, the hip and the knee showed major contributions to energy dissipation for double-leg landing, but for single-leg landing, the hip and the ankle were the dominant energy dissipaters Additionally, in the frontal plane, the hip acted as the key energy dissipater during double-leg landing, while the knee contributed the most energy dissipation during single-leg landing The findings collectively indicated that different energy-dissipating strategies were adopted for different landing techniques (double-leg and single-leg) in different planes (sagittal and frontal), which may imply that a certain lower extremity joint or pair of joints is more efficient in absorbing impact energy than others in a particular condition Therefore, a lack of sufficient energy dissipation from these key energy dissipaters due to improper execution of a double-leg or single-leg landing maneuver may increase the risk of sustaining lower extremity injuries

6.2 Investigation of Knee Injury Mechanisms

Equipped with the kinematics information from the prior motion analysis studies, ~70deg was identified as the knee flexion angle achieved at peak GRF during landing from large heights Therefore, the next step was to develop a test protocol to apply simulated landing impact to and induce ACL failure in knee specimens flexed

at 70deg In order to verify the feasibility of the test protocol, preliminary testing was conducted using porcine specimens We like to clarify that the use of porcine specimens was not to imply that pig knee responses to impact are a good model of human knee responses to impact The porcine model was chosen based on a prior work by Xerogeanes et al (1998), which demonstrated similar ACL load responses between porcine and human cadaveric specimens Furthermore, compared to human cadaveric specimens, the porcine model allowed better comparisons between non-impact (control) and impact test groups due to their relatively lower inter-specimen variability

The aim of this preliminary work in Stage B was to induce ACL failure in porcine knee specimens and examine the cartilage damage profile sustained during ACL failure from impact joint compression The most significant aspects were that (1) the specimen responses (Fz, θz and Dx) provided vital kinetics and kinematics data on the knee joint in the presence or absence of ACL failure during compression and (2) modified Mankin grading of the cartilage sections at various sites revealed the extent and distribution of impact damage during ACL failure from joint compression

The major limitation of this study was, perhaps, the use of a porcine model to investigate the ACL failure mechanism and the tibial cartilage damage profile Meyer and Haut (2005) showed that excessive joint compression on cadaveric tibiofemoral joints led to ACL failure The experimental setup was adapted in the current study,

which demonstrated that ACL failure also occurred in whole porcine knee specimens during excessive joint compression This clearly illustrated, to some extent, a similarity in ACL failure mechanism between the two models Furthermore, Xerogeanes et al (1998) suggested the porcine model as the preferred model for evaluating human ACL function, based on results that indicated close similarities in terms of the in-situ forces at the tested load range and knee flexion angles, and the load distribution between the ACL bundles for both models This further validated the choice of the porcine model as an appropriate preliminary option for investigating the cartilage damage profile during ACL failure

Another limitation was the absence of muscle forces during joint compression Pflum et al (2004) and Urabe et al (2005) have reported dominant quadriceps activity during landing, which consequently enhanced ACL and joint contact loading However at 70-deg knee flexion, the line of action of the quadriceps was almost parallel to the long axis of the tibia (Herzog and Read, 1993), hence the resulting anterior quadriceps pull on the ACL would be minute compared to the anterior tibial shear component developed from the compressive force in this study While not having muscle activation may likely affect posterior femoral displacement and tibial rotation, this simplified experiment was able to subject the specimens to a standardized set of loading conditions and demonstrate that significant joint compression, without muscle contraction, can solely induce ACL failure Due to the omission of the quadriceps contraction, it is important to note that the results may present a conservative underestimation of the cartilage damage profile during actual landing

Another constraint of the analysis was the successive compression trials on each specimen, which potentially introduced micro-damages in the soft tissue

structures especially at the ACL as the ACL restrains the rotational and anterior translational motion of the tibia (Fetto and Marshall, 1979) The results obtained, therefore, would be more suitable for addressing ACL failure during consecutive landings Applying a single high-impact load to induce ACL failure would be difficult due to specimen variability and would not be tenable for future experiments involving cadaveric specimens which are usually not easily procured

Based on the specimen responses, the dominant factors leading to ACL failure were likely the axial tibial rotation and the posterior femoral translation which would create rotational and translational stresses on the ACL Several studies reported that anterior tibial load alone can induce relative posterior femoral displacement (Sakane

et al., 1997; DeMorat et al., 2004) while others demonstrated that axial tibial rotation can elevate relative posterior femoral displacement and ACL loading (Gabriel et al., 2004; Yamamoto et al., 2004) Primarily, the ACL failure mechanism upon joint compression involved the development of the anterior tibial shear forces due to the compressive force acting on the posterior tibial slope (Pflum et al., 2004), which directly increased relative posterior femoral displacement; the anterior tibial shear forces acting along the tibial plateau will also give rise to internal/external tibial rotation depending on the magnitude of the shear forces in each tibial compartment The combination of the axial tibial rotation and posterior translation would, therefore, enhance the risk of ACL failure

Interestingly, three different variations on the failure mechanism were noted in terms of the axial tibial rotation Precautions were implemented in ensuring the same preparation procedures for all specimens, specifically in the positioning of the bone segments (tibia and femur) of each specimen within the potting cups through the use

of a guide plate and clamps Therefore, the observed inconsistency in axial tibial

rotation during ACL failure would likely be due to specimen variations, especially the bone geometry In addition, we expect that the type of axial tibial rotation may be partially attributed to the presence of knee valgus or varus Although we have made efforts to ensure a neutral frontal plane angle by visual inspection prior to impact loading, a slight knee valgus or varus may still exist depending on the knee joint compliance We expect that a knee valgus will promote early lateral tibiofemoral contact, which translates into anterior tibial shear forces at the lateral tibial compartment and thus induces internal tibial rotation On the other hand, a knee varus will facilitate early medial tibiofemoral contact and increases anterior tibial shear forces at the medial tibial compartment, which results in external tibial rotation In the situation whereby tibiofemoral contact occurs at about the same time for both medial and lateral compartments, anterior tibial shear forces will develop at both compartments, which tend to lead to posterior femoral displacement with a relatively low axial tibial rotation

K1 and K2 indicated higher damage scores at the anterolateral (AL/EL) and posteromedial (EM/PM) tibial compartments This damage profile suggests a pivoting

of the lateral femoral condyle on the anterolateral tibial cartilage while relatively more sliding occurs on the medial femoral condyle along the medial tibial cartilage, which may have led to the external tibial rotation observed for these specimens For K5 and K6, their damage profiles revealed greater damage at the anteromedial (AM/EM) and posterolateral (EL/PL) compartments, which may imply that the medial femoral condyle acts as a pivot on the anteromedial tibial cartilage with sliding occurring at a larger extent at the lateral femoral condyle along the lateral tibial cartilage; hence this suggests a possible internal tibial rotation K4 did not show substantial difference in scores between medial and lateral tibial cartilages, which may have reflected a low

extent of tibial rotation at the time of peak compressive force during (+)ACL failure

It appeared that ACL failure in K4 was dominantly caused by the posterior femoral translation, rather than a synergistic effect with rotation, because the extent of sliding for both femoral condyles along the tibial cartilages may have been similar as marked

by the higher damage scores at the posterior region It was interesting to note that K3 had a higher damage at the IL site, unlike other specimens This observation indicated

a greater concentration of impact load near the tibial spine, which may have contributed to an excessive accumulation of load on the underlying cement support, thus breaking it The proximity of the lateral tibiofemoral contact point to the tibial spine may also result in impedence of the femoral condyles, manifesting in a lack of both axial tibial rotation and posterior femoral translation

Generally, the specimens that underwent ACL failure incurred an overall modified Mankin score of 3.57-4.16 Kleemann et al (2005) reported a mean score of 3.2 for human osteoarthritic cartilage at a mild degeneration stage and 5.7-7.6 for moderate to severe degeneration stages It is critical to note that direct comparison of these values to the current study is not tenable due to differences in specimen model However, it can be inferred that the impact loads obtained (which induced ACL failure in the porcine knee specimens) have rendered the cartilage to damage equivalent to that of early OA

Thambyah et al (2006) reported, based on cadaveric specimens, that the articular cartilage beneath the meniscus had a significantly higher stiffness and modulus (but lower cartilage thickness and subchondral bone content) than the articular cartilage not covered by the meniscus This previous study concluded that the articular cartilage beneath the meniscus would be less prepared for weight-bearing

as compared to the exposed articular cartilage In the current porcine study, the

specimens that underwent ACL failure indicated greater damage at the unexposed articular cartilage (anterior (A), exterior (E) and posterior (P) sites), especially significant at E site, than at the exposed articular cartilage (interior (I) site) Since impact forces are usually more concentrated at the exposed region (tibiofemoral contact point) than the meniscus-protected regions, the results suggested that the I sites had better resilience to impact during ACL failure from joint compression

Meachim (1976) found that the meniscus-covered posterior segment and exposed region of the tibial cartilage were more inclined towards conversion to extensive overt fibrillation and localized defect with increasing age The progression

of cartilage defects was reported recently to occur more frequently near the central region of the tibial plateau (Ding et al., 2007) These previous data were collected from the general population and hence strongly reflected the development and extent

of cartilage defects sustained during daily activities over time This study has shown that the impact forces during ACL failure would induce more damage at the periphery (meniscus-covered anterior, exterior and posterior segments) of both tibial compartments than the exposed interior region There is a strong possibility that the damage at the anterior segment (A site) was a result of the pivoting of the femoral condyle on the tibial compartment to allow for tibial rotation Damage at the exterior and posterior segments (E and P sites respectively) were present due to the sliding of the femoral condyles on the posterior tibial slopes during joint compression The pivoting and sliding motion introduced substantial tibial contact and shear forces, resulting in cartilage fibrillation and delamination at A, E and P sites However, for the exposed interior region (I site), while surface disruption was less frequent, the development of subchondral damage (tidemark disruption) during ACL failure may heal into stiffer constructions such as subchondral sclerosis (Johnson et al., 1998)

Alteration of the overlying hyaline cartilage, such as loss of physiological distribution

of proteoglycans, due to subchondral fracture has been reported previously (Lahm et al., 2004); this would further subject the cartilage region to increased risk of sustaining focal defects Therefore, based on the findings in this study, one would be likely predisposed to an accelerated risk of knee joint degeneration at the time of ACL injury

Additionally, the prevalence of bone bruise in patients with ACL injuries has been evaluated to be 70–80% (Zeiss et al., 2005) Bretlau et al (2002) reported that 62.5% of the patients with ACL injury had bone bruises during the 1st examination and these lesions were found to occur more frequently at the lateral tibial plateau Other studies have also indicated that the posterolateral tibial compartment tends to develop bone bruises after ACL injury (Murphy et al., 1992; Brandser et al., 1996)

On the other hand, Rosen at al (1991) illustrated that 40% of the patients who had an ACL injury, had also injuries to the posteromedial tibial plateau Speer et al (1995) further reported a 29% frequency of medial tibial plateau lesions in skiers with ACL tear The disparity between these studies suggests that the lesion locations at the tibial plateau may be dependent on the type of ACL injury mechanism involved In the current study, the sites with high damage scores (especially the posteromedial and posterolateral tibial compartments) were in line with the bruise locations previously reported, indicating the porcine model may undergo similar ACL injury mechanisms

as that of the human model during an impact compression of the knee joint

This study has verified the feasibility of the porcine model to assess the role of the human ACL during impact joint compression Considering the cost, size, age variability and availability of cadaveric specimens, investigators may wish to adopt the porcine model to evaluate their preliminary test protocols During joint

compression, the impact loads that result in ACL failure can also potentially inflict considerable tibial cartilage damage; the extent and distribution of cartilage damage may be affected by the type of ACL failure mechanism Hence, future strategies in ACL injury prevention, such as improvement of brace designs, should also examine and alleviate the damage profile of the cartilage during impact landing

Armed with the test protocol for applying simulated landing impact to porcine specimens, it was subsequently translated for use with cadaveric specimens Therefore, the next study in Stage B aimed to investigate the effect of landing impact loads on the induction of ACL failure, and the extent and distribution of tibiofemoral cartilage damage in cadaveric knees The study served 3 critical aspects: 1) a significant drop

in compressive force response was used as an indicator for major ACL failure because intact cadaveric knees were tested, 2) volume analysis, thickness measurement and histological techniques were employed to assess cartilage deformation and damage inflicted by repetitive incremental landing impact loads leading to ACL failure, and 3) modified Mankin scoring of cartilage regions illustrated the extent and distribution of damage

One constraint of this study was that muscle forces were not introduced during impact loading While the quadriceps is a dominantly active knee extensor during landing (Pflum-et-al.,-2004;-Urabe-et-al.,-2005), it also affects ACL loading depending on the status of knee flexion At 70-deg flexion (current-study), the line-of-action of patella tendon/quadriceps is oriented nearly parallel to the long tibial axis (Herzog-et-al.,-1993;-Zheng-et-al.,-1998); thus the resultant anterior quadriceps pull

on the ACL would be minuscule Though the absence of muscle activation may affect posterior femoral displacement and axial tibial rotation, the simplified landing impact test was capable of imposing standardized loading conditions upon the intact

specimens and established that repetitive impact loading of the knee joint, without muscle contraction, can exclusively induce ACL failure It is also crucial to understand that the cartilage damage profiles obtained will present a conservative underestimation compared to actual impact landings, due to exclusion of quadriceps force

We like to highlight that muscle forces play an important role for the propensity of knee injuries during landing Aune et al (1995) showed that hamstring contraction helps to resist anterior tibial shear force in rats and thus protect the ACL

In addition, DeMorat et al (2004) reported that aggressive quadriceps loading can produce substantial anterior tibial translation and induce ACL injury in human cadaveric knee specimens These studies indicated the contributions of the different joint muscles to ACL injury risk However, in the current study, we sought to investigate the sole influence of compressive impact forces to ACL failure, independent of muscle forces This would permit us to understand how the compressive impact forces can directly contribute to anterior tibial translation and axial tibial rotation, which are the key risk factors in the ACL failure mechanism We acknowledge the importance of muscle loading in simulating a more physiological landing impact, and have conducted pilot musculoskeletal modeling work using OpenSim (v1.8.1, Simtk.org, US) to estimate the magnitude of muscle forces present during landing from different heights However, these preliminary results are not within the scope of this dissertation

In the current study, the experimental setup was adapted from a previous work (Yeow et al., 2008), which constrained the mounted specimen to 70-deg flexion to simulate a landing posture Hewett et al (1996) reported that the knee may be flexed 60-80 deg during jump-landing while Devita and Skelly (1992) found that stiff-style

landing, compared to soft-style landing, generated a greater peak ground impact and a flexed knee position at impact, which averaged 77deg Hence, the knee posture of 70deg was chosen for the impact compression trials It should be acknowledged that this is a limitation to the study as it is not feasible to obtain the knee flexion angle of a subject at the instant when ACL injury is incurred upon landing

Another limitation was the successive compression trials conducted on each specimen, which can latently propagate micro-damages in soft tissue structures especially the ACL as its role is to provide rotational and anterior translational restraint on the tibia (Fetto-et-al.,-1979) Gradual weakening of ACL structure can be observed from the steady increase in Fdrop with progressive loading until the final drastic drop Applying a single known impact load to induce ACL failure in cadaveric specimens would be impractical due to specimen variability and difficulties of procurement in terms of age, condition and quantity Therefore, we have decided to perform repeated incremental loading so that we were able to apply sufficient impact forces to induce ACL failure in the specimens without bone fracture We did not conduct any pilots with repeated identical loadings as a reference for the damage accumulation effect, but the progressive drop in compressive force response over the series of impact trials indicated that micro-damages are likely to occur in the soft tissues of the knee joint with repeated impact loading (Fig 5.2.1B and 5.2.4C) Furthermore, the test protocol which allows incremental loading facilitated us in differentiating between pre-failure and ACL failure, and further ensures that the applied load is sufficient to induce a major ACL failure only, without causing bone fractures that complicate analysis Furthermore, this protocol allows us to investigate the effect of repetitive landing impact loads on the outcome of cartilage lesions in order to better understand repetitive knee stress injuries

In addition, we acknowledge that the presence of a forward velocity is common during landing and can contribute to an anterior-posterior motion component, which further influence the propensity for ACL injury However, it have been previously reported that an external posterior force would be exerted onto the lower extremities by decelerating the body’s forward momentum during landing; this posterior force further acts on the tibia to reduce ACL strain (Shin et al., 2007) Moreover, Pflum et al (2004) noted that the ground reaction can apply a posterior shear force to the lower leg during landing, which acts to unload the ACL These studies suggested that posterior shear forces that arise in a landing scenario with forward velocity would unlikely increase the risk for ACL injury Therefore, in the current study where the cadaveric specimens were subjected to a simulated landing impact with negligible forward velocity, we were able to assess the direct effect of impact compressive forces on the anterior-posterior tibiofemoral joint motions and consequently the outcome of ACL failure

It should be noted that meniscal damage is also one of the common knee injuries incurred during landing-intensive sports (Ferretti et al., 1990; Zazulak et al., 2007) However, this form of damage is not assessed in the current study as the main focus of the study was to investigate how the ACL failure mechanism can influence the extent and distribution of cartilage damage It is possible to assess the extent of meniscal deformation based on the pre- and post-impact MR images for all the cadaveric specimens, but this prospective work would be conducted outside this dissertation

The specimen responses during impact loading indicate that ACL failure was mainly attributed to posterior femoral displacement and axial tibial rotation, which predispose ACL to translational and rotational strains Anterior tibial load and/or axial

tibial rotation can induce relative posterior femoral displacement and elevate ACL stress (DeMorat-et-al.,-2004;Gabriel-et-al.,-2004;Sakane-et-al.,-1997) Upon impact, the compressive force acting on the posterior tibial slope initiates anterior tibial shear forces (Pflum-et-al.,-2004), which directly increase posterior femoral displacement and produce internal/external tibial rotation depending on the magnitude of shear forces in each compartment This synergy between anterior tibial translation and axial tibial rotation would consequently augment ACL failure risk

In addition, menisci-covered cartilage possesses lesser subchondral bone content & cartilage thickness than exposed cartilage (Thambyah-et-al.,-2006) These findings implied that menisci-covered cartilage would be less primed for weight-bearing compared to exposed cartilage In this study, the cartilage damage profiles generally indicated greater damage around menisci-covered regions for H1-H4 than exposed region; however for H5, the highest damage was located at exposed regions than neighboring covered regions In fact, impact forces are usually more concentrated at exposed regions, where the tibiofemoral contact point exists Therefore, the results suggest that exposed regions generally possess superior resilience to impact (relative-to-covered-regions), but if these regions are subjected to excessively large impact loads, they may present higher damage than their protected neighbors

Several studies found that menisci-covered posterior and exposed interior tibial cartilage regions were more inclined towards extensive fibrillation and localized defect with increasing age (Ding-et-al.,-2007;-Meachim-et-al.,-1976) Additionally, Yoshioka-et-al.-(2004) reported cartilage thinning at posterior regions of femoral condyles in most OA patients, based on MR imaging, while anterior regions of tibiofemoral cartilages exhibited lesser degradation (Biswal-et-al.,-2002) These

studies are indicatory of the progression and severity of cartilage defects developed during daily activities In the current study, though there was no significant difference

in average specimen age between control and impacted groups, it should be acknowledged that the specimen age may be a confounding factor that will influence the histological results

The results further indicated that repeated incremental landing impact loads leading to ACL failure were able to induce high levels of damage across tibiofemoral cartilages in impacted specimens, compared to non-impacted controls Moreover, cartilage deformation was significant in both compartments, especially at the anterior and posterior tibial regions, and the exterior, posterior and interior femoral regions This could be attributed to the posterior shift of tibiofemoral contact during impact loading; this sliding motion of the femur, relative to tibia, impose considerable tibiofemoral contact and shear forces, resulting in cartilage fibrillation and delamination at the anterior, exterior and posterior tibial regions, and the exterior, posterior and interior femoral regions The presence of tidemark disruption in most regions will increase risk of sustaining focal defects; this subchondral damage may lead to sclerosis (Johnson-et-al.,-1998) and loss of physiological proteoglycan distribution in the overlying cartilage (Lahm-et-al.,-2004)

Bone bruises are highly rampant in patients following ACL injuries (Zeiss-etal.,-1995) Some studies reported that these lesions occur frequently at posterolateral and posteromedial tibial compartments (Brandser-et-al.,-1996;-Bretlau-et-al.,-2002;-Rosen-et-al.,-1991;-Speer-et-al.,-1995) For the femoral counterparts, they usually develop around middle and posterior condylar regions (Fayad-et-al.,-2003;-Sanders-

-et-al.,-2000) Though bone bruise locations were termed as footprints of ACL failure mechanism (Fayad-et-al.,-2003;-Hayes-et-al.,-2000), the actual tibiofemoral motion

present during clinical ACL failures with bone bruising is largely uncertain In this study, there were high damage scores at the exterior, posterior and interior femoral regions, with substantial deformation at both tibiofemoral compartments These findings indicated that the specimens in the study underwent an ACL failure mechanism based dominantly on anterior tibial translation and were consistent with the relative posterior femoral displacement profiles and values obtained, and the actual tibiofemoral motion during ACL failure

The results collectively illustrated that substantial tibiofemoral cartilage deformation and damage can be attributed to the repeated incremental landing impact loads that lead to ACL failure These loads can also inflict a damage profile that reflects a failure mechanism governed by anterior tibial translation Hence, an individual may be potentially predisposed to an accelerated risk of knee joint degeneration when performing high-risk landing tasks that can latently lead to ACL injury Though the findings only revealed that the repeated application of incremental landing impact loads to intact male cadaveric specimens can induce ACL failure and tibiofemoral cartilage deformation and damage, there would likely be some differences in these outcomes for female specimens Previous studies have shown that women displayed different ACL ultrastructure and biomechanical properties (Hashemi-et-al.,-2008), and greater knee valgus during landing (Russell-et-al.,-2006), compared to men These biomechanical differences are important risk factors for ACL failure Therefore, the application of this impact protocol to female cadaveric specimens in a physiological valgus position may potentially result in the occurrence

of ACL failure at lower impact loads, and greater deformation and damage at lateral tibiofemoral cartilage compartments Future studies should consider testing of female cadaveric specimens, and inclusion of muscle forces and knee valgus/varus to

evaluate their effects on the knee’s propensity for ACL failure, and the extent and distribution of cartilage damage

Apart from biomechanical factors, it is important to note that gender differences in ACL injury incidence can be due to hormonal and tissue compositional factors Komatsuda et al (2006) found that the ultimate tensile stress and linear stiffness of the ACL were lower in rabbits with high serum estrogen levels More recently, Yoshida et al (2009) illustrated that ovarian hormones can alter the extracellular matrix composition of the ACL Moreover, the COL5A1 gene, which encodes for the alpha1 chain of type V collagen, was demonstrated to be associated with increased risk of anterior cruciate ligament ruptures in female participants (Posthumus et al., 2009) These studies showed that hormones and tissue composition are part of the multi-factorial pathogenesis of ACL rupture in the female gender

Furthermore, it should be underscored that the frequency or habituality of impact may be an important influence factor with regards to landing-related studies Comparing person A with daily experience in landing maneuvers to person B with little experience in landing, we would expect that person A would be relatively more capable of efficiently coordinating his/her lower extremity joint motions to optimize energy dissipation during landing from different heights or using different techniques For the landing studies performed (Stage A), we have chosen recreational athletes as they have fairly similar experiences in performing landing maneuvers; they neither have superior landing abilities like competitive athletes, nor negligible landing experience Therefore, these subjects possess comparable landing strategies, which allowed us to examine the direct effects of landing height and technique on the biomechanical parameters that influence lower extremity injury risk during landing

However, in the event of excessively high GRF during repetitive landing, person A would be more vulnerable to damage accumulation in the soft tissues surrounding the lower extremity joint, which would likely lead to overuse injuries; person B would be able to recover relatively faster even after such a large landing impact due to the absence of damage accumulation and the rare exposure to landing impacts Thus, we expect that in our cadaveric impact studies (Stage B), the specimens that underwent repetitive impact loading would potentially sustain greater damage than the control specimens that did not experience loading Although we do not have prior knowledge whether the donors of the cadaveric specimens have different experiences, in terms of the frequency or habituality of performing landing maneuvers, we have checked the specimens for visible gross cartilage defects using

MR imaging with advice from the radiologist Moreover, the knee specimens were randomly allocated into control and impact groups Therefore, it is probable that the repetitive application of impact loads to the specimens can inflict substantial cartilage damage, as compared to the non-impact control specimens

While important data such as peak compressive, relative anterior tibial translation and axial tibial rotation were obtained from the prior two impact studies using the test protocol, it was not possible to measure the peak ACL and cartilage stresses during simulated landing impact with the knee specimens in their intact form Hence, the next objective of Stage B was to develop a finite element model of the tibiofemoral joint, validate it with experimental results, and use it to estimate peak ACL and tibial cartilage stresses during a simulated landing impact The important aspects of this study were: 1) the finite element model was validated with experimental results in terms of compressive force, posterior femoral displacement and axial tibial rotation, and 2) the finite element model predicted the extent and

distribution of ACL and tibial cartilage stresses during different magnitudes of simulated landing impact

The geometry of the cadaveric specimen tested in the impact protocol was adopted in the finite element model via reconstruction of the pre-impact MRI scans Additionally, the loading condition applied to the finite element model was based on the actuator displacement curve (6-mm amplitude at single 10-Hz haversine) introduced in the experimental setup The same experimental boundary conditions were used in the finite element model; specifically, axial rotation and axial displacement of the tibia, and anterior-posterior and medial-lateral displacements of the femur were permitted Moreover, the time profiles for compressive force, posterior femoral displacement (relative anterior tibial translation) and axial tibial rotation determined at 6-mm actuator displacement in the finite element model were fairly consistent with the results obtained from the experiment Therefore, the findings reasonably verified that a validated finite element model was developed based on the

simulated landing impact condition

From the model, the peak compressive force sustained by the cadaveric knee increased with elevation in actuator displacement The increase in peak compressive force also demonstrated a corresponding aggravation of posterior femoral displacement with reduced axial tibial rotation angle during the course of impact The findings suggested that a high peak compressive force during a landing impact promotes large anterior tibial translation, which can thus increase ACL stress; the range of peak ACL stresses obtained fell well within a previous study that applied a combination of 1.15-kN compressive load and 0.13-kN anterior tibial load (Pena et al., 2006) These increases in anterior tibial translation and ACL stress with higher peak compressive force demonstrated that the ACL injury risk will be potentially

heightened when one sustained a large landing impact Furthermore, the middle region of the ACL incurred the greatest stress, which can potentially lead to a mid-substance tear; previous studies that conducted simulated landing impact studies have also noted the presence of mid-substance ACL ruptures in cadaveric specimens (Meyer et al., 2005; Yeow et al., 2009c)

Apart from exacerbated anterior tibial translation, a reduction in external tibial rotation was noted during early impact phase and a presence of internal tibial rotation with increased peak compressive force These results illustrated that increased compressive loading at the knee joint tends to encourage the tibia to undergo internal rotation As the ACL functions as a secondary restraint to internal tibial rotation (West and Harner et al., 2005), the presence of internal tibial rotation during simulated landing impact is likely to contribute, in part, to greater ACL stress; moreover, the application of an internal tibial torque was reported to be able to induce

a rapid increase in anterior tibial translation (Kanamori et al., 2002) Altogether, a high peak compressive force can induce substantial anterior tibial translation and internal tibial rotation, which can synergistically contribute to an elevated peak ACL stress, and thus potential ACL failure

Besides an increased risk of ACL failure, it was observed that a high peak compressive force can also latently induce considerable stresses on the tibial cartilage, particularly at the posterior region of the medial compartment The large peak cartilage stresses of 16.6-23.0MPa observed at the medial tibial compartment for the three actuator displacement conditions was in the range of 14-35MPa that were previously reported to induce chondrocyte death and matrix damage (Adams, 2006; Yeow et al., 2009a) Additionally, it was reported that approximately 80-90% of people suffering from OA had cartilage defects in the medial compartment (Ding et

al., 2007) Furthermore, the presence of high compressive stresses at the posterior regions of both medial and lateral tibial compartments indicated the posterior shift of the tibiofemoral contact point due to the posterior femoral displacement induced by the compressive force, which is therefore likely to inflict substantial damage at the posterior tibial cartilage regions (Yeow et al., 2009c)

It should be noted that although the medial cartilage compartment exhibited greater compressive stresses compared to the lateral compartment in all three different simulations of this FE model, the prior experimental study did not show a substantial difference in Mankin scores between the compartments We expect that the repeated application of incremental landing impact loads to the human cadaveric knee specimens in the experimental study could have introduced notable levels of damage

to both cartilage compartments, wherein the Mankin grading system may not be adequate in differentiating the extent of damage between the compartments

We acknowledge that h-convergence testing may be useful to justify the number of elements used in the knee FE model However, it has been reported that mesh refinement is more penalizing in explicit than in implicit techniques since it affects the minimum time step Therefore, we developed a FE model with a range of element sizes (1-5mm) that is compatible with a minimal time step of the order of 1µs and is suitable for studying the kinematics and global response at the joint level (Ulrich et al., 1998; Beillas et al., 2004) Nonetheless, local refinement of the mesh could be used in order to study highly localized phenomena (such as stresses at ligament insertion points) without penalizing the overall computation time In the current study, we sought to examine the gross responses of the soft tissues, especially the ACL and articular cartilage While our simple FE model was not able to accurately replicate experimental behavior, it displayed moderate similarity with the

experimental results and allowed a fair estimation of the impact stresses that may be sustained by the ACL and tibial cartilage

In summary, a finite element model of the tibiofemoral joint was constructed and validated with experimental results obtained in a prior simulated landing impact experiment This model was then used to estimate the effect of increased compressive force on relative anterior tibial translation and axial tibial rotation, and to estimate the extent and distribution of ACL and tibial cartilage stresses during simulated landing impact The findings collectively illustrated that the increase in peak compressive force aggravated anterior tibial translation and internal tibial rotation, and further led

to elevated tensile stresses at the middle region of the ACL that would heighten the risk of mid-substance failure Moreover, the large peak compressive forces were found to inflict substantial cartilage stresses, especially at the posterior regions of both medial and lateral tibial compartment, which will exacerbate cartilage damage and accelerate the risk of developing post-traumatic OA

With the knowledge of the peak compressive stresses that were sustained by the cartilage during simulated landing impact, the next step was therefore to understand how these stresses may lead to cartilage damage or even degeneration Hence, this next study in Stage B aimed to examine damage and degenerative changes

at the osteochondral level upon a single simulated landing impact The most significant aspects were that (1) menisci-covered and exposed explants exhibited considerable differences in load response when subjected to the same extent of impact compression (the menisci were not involved during impact testing); and (2) the post-impact assessment of damage and degenerative changes allows us to understand the individual capacity of menisci-covered and exposed regions in resisting impact load

Previous studies demonstrated impact stresses of 20-30MPa could damage both the cartilage matrix (D’Lima et al., 2001; Milentijevic eta l., 2005) and the underlying bone (Haut, 1989) In this study, mean peak compressive stresses of 20.2[7.6]MPa and 34.9[12.1]MPa were observed for menisci-covered and exposed explants respectively when subjected to 2-mm displacement compression At these peak impact stresses, there were significant structural damages such as matrix disruption at superficial and intermediate zones, and tidemark lesions, which are consistent with previous studies (Torzilli et al., 1999; Ewers eta l., 2001) Degenerative changes that arose as a result of large impact stresses in the current study were fairly in line with past reports (Mrosek et al., 2006; Loening et al., 2000; Morel et al., 2006) For instance, a gradual elevation in cartilage volume post-impact and a reduced GAG content up to Day 14 were noted A previous report by Loening

et al (2000) indicated that cartilage swelling and GAG loss were present up to 7 days post-impact A substantial presence of surface cartilage cell death was also found upon impact and the extent of cell death persisted up to 14 days Morel et al (2006) reported an induction of cell death in the superficial zone upon impact stress, which subsequently spread to deeper zones up to 11 days Furthermore, Mrosek et al (2006) illustrated a notable decrease in type II collagen and aggrecan expressions in the cartilage 6 months post-impact An increase in type I collagen expression was also observed in the previous study However no substantial type I collagen staining was detected in the study at Day 14, which suggests that post-traumatic type I collagen production by chondrocytes may only be significant at much later time-points Moreover, the microCT scans revealed a visible change in surface cartilage morphology over time for pre- and post-impact explants, which was reflective of impact damage, cartilage swelling and surface degradation

Most importantly, the study has indicated a clear difference between the load responses of menisci-covered and exposed explants when subjected to the same 2-mm displacement compression Previously, Thambyah et al (2006) showed that exposed regions possess a significantly larger cartilage thickness and subchondral bone density compared to menisci-covered regions These factors may contribute in part to a higher overall stiffness in exposed osteochondral explants, therefore these explants will experience a greater peak compressive impact stress when compressed at 2-mm displacement, relative to menisci-covered explants; this further emphasized the superior load-bearing capacity of the exposed regions

Moreover, no major differences were detected between menisci-covered and exposed impacted explants in terms of cell death, GAG and collagen content, and Mankin scores This showed that though exposed explants generally sustained a higher impact load and the same magnitude of compressive deformation, the damage and degenerative changes following impact was not found to be significantly different compared to menisci-covered explants This is strongly indicative of the innate biomechanical property of exposed osteochondral regions in resisting trauma-induced damage and degeneration

However, a significant elevation in cartilage volume was noted for exposed explants, which implied cartilage swelling This swelling may be attributed to the diminished capability of the disrupted superficial collagen network to counteract proteoglycan-induced swelling pressure (Loening et al., 2000; Bonassar et al., 1995) Loening et al (2000) related the significant decrease in equilibrium and dynamic stiffness of bovine explants after injurious compression to collagen network damage; the extent of cartilage swelling was also found to increase with the severity of injurious load Moreover, Bonassar et al (1995) reported that increase in cartilage

swelling concomitantly occurred with degradation in type II collagen The data appeared to suggest that the lower compressive load experienced by the menisci-covered explants could lead to a lesser extent of collagen network damage, hence less swelling when compared to the exposed explants Nevertheless, additional studies would need to be performed in order to better understand the relationship between compressive load magnitude, collagen network damage and cartilage swelling.

The single 2-mm displacement compression or an equivalent simulated landing impact of 20.2-34.9MPa applied to the osteochondral explants was able to induce damage and degeneration with an overall Mankin score of 6.3-7.1 at Day 14 (compared to control score of 2.3-2.7) Kleemann et al (2005) reported a mean Mankin score of 3.2 for human cartilage at mild OA stage and 5.7-7.6 for moderate-severe stages Though a porcine model was used in this study, the results demonstrated that the landing impact compression introduced to the explants was sufficient to inflict osteochondral damage and degeneration and present at least a moderate OA status

The main limitation of this study was the use of a porcine model to investigate the post-impact osteochondral damage profile As fresh cadaveric specimens are usually not easily procured and are often derived from aged donors, the use of these specimens as controls and for impact testing would be difficult for attaining reliable results and interpretation In the current study, the porcine model was used in place to minimize sample variability as they were derived from a common source

Furthermore, the daily activities performed by pigs are considerably more sedentary compared to humans, hence their knees do not normally sustain high impact

or have a history of knee injury Therefore, the porcine model would be appropriate as control and impacted specimens It is also important to note that the juvenile porcine

model used in this study may be structurally different from skeletally mature pigs or adult human model A complex laminar collagen network was seen in juvenile bovine cartilage whereas more mature tissue exhibited a simpler network pattern (Olivier et al., 2001) Furthermore, the structure of mature human cartilage possessed three distinct collagenous zones, while juvenile animal tissue displayed three, five or seven laminae/zones (Nissi et al., 2006) Though juvenile animal tissue is widely used as a model for mature human cartilage, their structural variations should be recognized The results obtained in the current study for juvenile porcine model may not be directly translatable to mature porcine/human models, but the results underscored the differences in mechanical load responses between menisci-covered and exposed osteochondral regions, which are also present in mature human cartilage (Thambyah

et al., 2006) These differences are relevant for understanding the susceptibility to post-traumatic damage and degeneration between the different osteochondral regions

Another constraint was the use of culture medium to maintain the viability of osteochondral explants Though it would be most ideal to use the synovial fluid to better simulate the physiological condition of the knee joint during a landing impact, the synovial fluid is quite limited in quantity and infeasible to replace unlike the regular culture medium change Obtaining a synovial fluid substitute would be difficult due to its complex composition which includes hyaluronan and lubricin (Jay

et al., 2007; Meng and Long, 2008) Moreover, the cytokine levels in the synovial fluid can vary among normal, injury or osteoarthritic conditions (Elsaid et al., 2008; Richette et al., 2008) For instance, the synovial fluid can have an elevated level of pro-inflammatory cytokines, such as interleukin-I, which is known to suppress type II collagen synthesis and promote type I collagen production (Goldring et al., 1988) This may contribute in part to the lack of type I collagen staining in the

immunohistochemical results However, due to the many components that exist within the synovial fluid which may complicate the analysis, it would be more appropriate to adopt the standard culture medium as a simplified substitute, with appropriate osmolarity (280mOsm), pH (7.4) and supplement concentrations, to maintain tissue viability so that the direct effect of landing impact compression on osteochondral damage and degeneration, independent of the synovial fluid can be investigated

It should be recognized that osmolarity may affect the chondrocyte volume and sensitivity to load (Bush et al., 2005) while pH may influence chondrocyte viability (Dontchos et al., 2008), thus the standard culture medium was adopted as it has been a common approach for previous related studies (Morel et al., 2006; Patwari

et al., 2004; Bush et al., 2005) It should be acknowledged that the culture medium does not reflect the extracellular environment in-situ and, therefore, may not be conducive for normal chondrocyte function However, all test groups were subjected

to the same composition of culture medium, hence any differences between test groups would likely be due to the applied compressive load and the inherent mechanical properties of the osteochondral regions

Altogether, these findings have indicated that exposed osteochondral regions possessed superior load-bearing capacity and better resilience to damage and degeneration compared to their menisci-covered counterparts Exposed regions can sustain relatively larger compressive impact stresses before sustaining the extent of damage and degeneration that is comparable to menisci-covered regions This property is especially important during daily locomotion as the exposed region normally acts as the tibiofemoral contact area and is consistently subjected to greater joint loading On the other hand, meniscectomy or damaged menisci will however

‘expose’ menisci-covered regions and increase the risk of damage and degeneration in

these areas The findings have shown that these regions were not only susceptible to impact-induced damage, they were also more inclined towards post-traumatic degeneration Apart from the altered knee joint biomechanics arising from meniscectomy (Sturnieks et al., 2008), the inferior resistance of menisci-covered regions, against impact-induced damage and degeneration, is a potential factor that may contribute in part to the meniscectomy model of OA

The previous study was able to identify the extent of osteochondral damage after simulated landing impact, but the status of lesion is unknown at the point of peak impact compression, yet this information is essential for understanding the true damage inflicted by the impact Thus, the purpose of the next study in Stage B was to examine the differences in the extent of damage and deformation incurred at the tibial cartilage between the end of the loading phase (at peak displacement compression of the tibial cartilage during a simulated landing impact) and the end of the unloading phase (removal of load after the simulated landing impact)

The key aspects of this study were: 1) the extent of cartilage damage was substantially lower in sustained compression condition than in non-sustained compression condition, 2) the total cartilage thickness was markedly reduced in sustained compression condition compared to non-impact control and non-sustained compression condition, and 3) the superficial zone thickness was notably lower in both sustained and non-sustained compression conditions, relative to the non-impact control

The peak impact stresses incurred by the various regions was well within the range of impact stresses (15-30MPa) previously reported to inflict substantial cartilage damage (D’Lima et al., 2001; Haut, 1989, Loening et al., 2000; Milentijevic

et al., 2005) Although the interior cartilage regions in both medial and lateral