Knee biomechanics during impact landing understanding injury mechanisms and developing prevention strategies 2

The peak GRF was achieved within 20% of the landing phase; similar observations were noted for peak knee flexion angular velocity and knee joint power.. 5.1.1: Representative profiles of

CHAPTER 5 RESULTS

~ They say any landing you can walk away from is a good one ~

Alan Shepard

70

5.1 STAGE A – Understanding Biomechanics of Landing

5.1.1 Regression relationships of knee kinematics, kinetics and energetics with landing height

a Time profiles of knee kinematics, kinetics and energetics

Representative GRF data indicated the typical two-peak profile commonly observed during landing (Fig 5.1.1) The peak GRF was achieved within 20% of the landing phase; similar observations were noted for peak knee flexion angular velocity and knee joint power

Fig 5.1.1: Representative profiles of resultant ground reaction force (GRF), knee flexion angle, knee

flexion angular velocity and knee joint power during landing phase

b Peak ground reaction force

Between peak resultant GRF and landing height, a strong exponential regression relationship (R2=0.90-0.99, p<0.001; power=0.987-0.998) was noted for all the 5 subjects (Fig 5.1.2)

Fig 5.1.2 Exponential regression relationships of landing height with peak resultant GRF for all

subjects generally followed a y=ae bx equation, where y=peak resultant GRF, x=landing height, a,b=regression coefficients

c Knee flexion angle

The knee flexion angle at initial contact had an inverse-exponential regression relationship (R2=0.81-0.99, p<0.001 to p=0.006; power=0.834-0.978) with landing height (Fig 5.1.3A) Similarly, the knee flexion angle at peak GRF also possessed an inverse-exponential relationship (R2=0.84-0.97, p<0.001 to p=0.004; power=0.873-0.999) with landing height (Fig 5.1.3B)

72

Fig 5.1.3: Inverse-exponential regression relationships of landing height with knee flexion angles (A)

at initial contact and (B) at peak GRF for all subjects generally followed a y=aln(x)+b equation, where

y=knee flexion angles at initial contact/at peak GRF, x=landing height, a,b=regression coefficients

d Knee flexion angular velocity

In addition, non-linear regression between peak knee flexion angular velocity and landing height revealed an inverse-exponential relationship (R2=0.86-0.96, p<0.001; power=0.935-0.994) (Fig 5.1.4)

Fig 5.1.4: Inverse-exponential regression relationships of landing height with peak knee flexion

angular velocities for all subjects generally followed a y=aln(x)+b equation, where y=peak knee

flexion angular velocities, x=landing height, a,b=regression coefficients

e Knee joint power

Peak knee joint power demonstrated a substantial linear relationship (R21.00, p<0.001; power=0.990-1.000) with landing height (Fig 5.1.5)

=0.98-Fig 5.1.5: Simple linear regression relationships of landing height with peak knee joint power for all

subjects generally followed a y=ax+b equation, where y=peak knee joint power, x=landing height,

a,b=regression coefficients

f Regression relationships

The peak resultant GRF during landing typically followed an exponential regression relationship with landing height in the form of y=aebx, where y=GRF, x=landing height; the values for regression coefficients, a and b, ranged from 1.19 to 1.52 and 0.80 to 1.13 respectively (Table 5.1.1)

The knee flexion angles at initial contact and at peak GRF generally adopted

an inverse-exponential (natural logarithmic) regression relationship with landing height in the form of y=aln(x)+b, where y=knee flexion angle, x=landing height) For the knee flexion angles at initial contact, the values for a and b ranged from 2.70 to 10.81 and 32.34 to 40.78 respectively Regarding knee flexion angles at peak GRF, the values for a and b ranged from 2.97 to 12.52 and 64.10 to 67.38 respectively

74

In terms of knee flexion angular velocity, the peak value increased in an inverse-exponential regression relationship with landing height, where a and b ranged from 2.40 to 6.27 and 12.38 to 16.59 respectively For knee joint power, the peak value followed a strong linear regression relationship, y=ax+b, where y=joint power and x=landing height The a and b values ranged from -54.04 to -27.59 and -2.36 to 2.01 respectively

Table 5.1.1: Typical non-linear/linear regression equations and corresponding regression coefficients

obtained for Subject 1 with regards to the dependent variables: peak resultant GRF, knee flexion angles (at initial contact and peak GRF), peak knee flexion angular velocities and peak knee joint power A

R2-value near 1 indicates a strong non-linear/linear regression relationship between the independent (landing height) and dependent variables A p-value less than 0.05 implies that the independent variable can be used to predict the dependent variables

Regression Coefficients Dependent Variables Regression Equations

Knee Flexion Angle at

Knee Flexion Angle at

Peak Knee Flexion

Peak Knee

5.1.2 Effect of landing height and technique on sagittal knee joint kinematics and energetics

a Time profiles

The landing phase was denoted by the duration of landing between initial contact and maximum knee flexion GRF, knee flexion angular velocity and knee joint power peaked at approximately 25% landing phase while the knee flexion angle plateau-ed towards the end of landing phase (Fig 5.1.6)

Fig 5.1.6: Representative profiles of a subject for resultant ground reaction forces (GRF), knee flexion

angles, knee flexion angular velocities and knee joint powers during the landing phases of single-leg and double-leg landing

b Peak ground reaction force

A major increase (p<0.05) in peak resultant GRF was observed during leg landing from 0.6-m height, compared to 0.3-m height; single-leg landing was also

single-76

found to have significantly greater peak GRF (p<0.001) than double-leg landing at both heights (Fig 5.1.7)

Fig 5.1.7: Peak resultant ground reaction forces (GRF) during the landing phase of double-leg landing

and single-leg landing from 0.3-m and 0.6-m heights

* significant difference compared to double-leg landing at 0.3-m height (p<0.001)

** significant difference compared to single-leg landing at 0.6-m height (p<0.05) and double-leg landing at both 0.3-m and 0.6-m heights (p<0.001)

c Knee flexion angle

A significant elevation in knee flexion angles were noted at peak GRF and at maximum knee flexion during double-leg landing from 0.6-m height, relative to 0.3-m height The knee flexion angles were consistently greater (p<0.005) during double-leg landing than single-leg landing at all three events from both landing heights (Fig

5.1.8A and 5.1.8B)

Fig 5.1.8: Knee flexion angles at initial contact, at peak GRF and at maximum knee flexion during the

landing phase of (A) double-leg landing and (B) single-leg landing

* significant difference between 0.3-m and 0.6-m heights (p<0.05)

# significant difference compared to single-leg landing (p<0.005)

d Knee flexion angular velocity

The knee flexion angular velocity was the highest (p<0.01) during double-leg landing from the 0.6-m height; double-leg landing from 0.3-m height has a substantially greater knee flexion angular velocity (p<0.05) than the single-leg landing from 0.3-m height (Fig 5.1.9)

78

Fig 5.1.9: Knee flexion angular velocity at peak GRF during the landing phase of double-leg landing

and single-leg landing from 0.3-m and 0.6-m heights

* significant difference compared to single-leg landing at 0.3-m height (p<0.05)

** significant difference compared to double-leg landing at 0.3-m height (p<0.01) and single-leg landing at both 0.3-m and 0.6-m heights (p<0.001)

e Knee joint power

Similarly, the negative knee joint power was the largest (p<0.001) during double-leg landing from the 0.6-m height; double-leg landing from 0.3-m height delivered a higher negative knee joint power (p<0.05) than the single-leg landing

from 0.3-m height (Fig 5.1.10)

Fig 5.1.10: Knee joint power during the landing phase of double-leg landing and single-leg landing

from 0.3-m and 0.6-m heights

* significant difference compared to single-leg landing at 0.3-m height (p<0.05)

** significant difference compared to double-leg landing at 0.3-m height and single-leg landing at both 0.3-m and 0.6-m heights (p<0.001)

f Eccentric work

The eccentric work done during double-leg landing from the 0.6-m height was substantially larger (p<0.001), relative to that during double-leg landing from 0.3-m

height and single-leg landing from both heights (Fig 5.1.11)

Moreover, there was no statistically significant interaction (p=0.103-0.852) between landing technique and landing height for all parameters measured in this study The effect of landing technique did not depend on the level of landing height tested in the current study

Fig 5.1.11: Eccentric work done during the landing phase of double-leg landing and single-leg landing

from 0.3-m and 0.6-m heights

** significant difference compared to double-leg landing at 0.3-m height and single-leg landing at both 0.3-m and 0.6-m heights (p<0.001)

80

5.1.3 Differences in energy-dissipating strategies between landing techniques in both sagittal and frontal planes

a Joint angle at initial contact

Upon initial contact, the hip and knee joints adopted significantly greater flexion angles (p<0.05) than the ankle joint for both double-leg and single-leg landing (Fig 5.1.12A) The knee joint displayed a larger flexion angle at initial contact during double-leg landing than single-leg landing In the frontal plane, the hip joint exhibited more adduction at initial contact (p<0.05), compared to the knee and ankle joints, for both landing techniques (Fig 5.1.12B) The knee joint showed greater abduction at initial contact (p<0.05) than the ankle joint for both types of landing

Fig 5.1.12: Comparison of hip, knee and ankle joint angles at initial contact in the (A) sagittal and (B)

frontal planes during double-leg and single-leg landing

^ significant difference compared to single-leg landing (p<0.05)

* significant difference compared to ankle joint (p<0.05)

** significant difference compared to other two joints (p<0.05)

(A)

(B)

b Peak joint angle

The knee joint achieved a larger peak flexion (p<0.05) than the hip and ankle joints (Fig 5.1.13A) for both landing techniques The hip joint also attained a notably greater peak flexion (p<0.05) than the ankle joint Both the hip and knee joints demonstrated elevated peak flexion (p<0.05) during double-leg landing, relative to single-leg landing For the frontal plane, the hip joint indicated more peak adduction (p<0.05) than the ankle joint for both landing techniques (Fig 5.1.13B) The knee showed greater peak abduction (p<0.05) than the ankle joint during double-leg landing, and both the hip and ankle joints during single-leg landing

Fig 5.1.13: Comparison of peak hip, knee and ankle joint angles in the (A) sagittal and (B) frontal

planes during double-leg and single-leg landing

^ significant difference compared to single-leg landing (p<0.05)

* significant difference compared to ankle joint (p<0.05)

** significant difference compared to other two joints (p<0.05)

(A)

(B)

82

c Joint angular velocity

In the sagittal plane, the knee joint displayed greater flexion angular velocity (p<0.05) than the hip and ankle joints during double-leg, and the hip joint during single-leg landing (Fig 5.1.14A) The ankle joint showed markedly higher flexion angular velocity (p<0.05) than the hip joint for both landing techniques The hip and knee joints also exhibited larger flexion angular velocity (p<0.05) during double-leg landing than single-leg landing At the frontal plane, the knee achieved a higher abduction angular velocity (p<0.05) than the hip and ankle joints for both landing techniques (Fig 5.1.14B) The hip joint illustrated larger adduction angular velocity (p<0.05) than the ankle joint during single-leg landing

Fig 5.1.14: Comparison of peak hip, knee and ankle joint angular velocities in the (A) sagittal and (B)

frontal planes during double-leg and single-leg landing

^ significant difference compared to single-leg landing (p<0.05)

# significance difference compared to double-leg landing (p<0.05)

* significant difference compared to hip joint (p<0.05)

+ significant difference compared to ankle joint (p<0.05)

** significant difference compared to other two joints (p<0.05)

(A)

(B)

d Joint moments

For joint moments, the hip joint displayed larger peak flexion moments (p<0.05) than the knee and ankle joints for both double-leg and single-leg landing (Fig 5.1.15A) The ankle joint showed greater peak flexion moment (p<0.05) than the knee joint during single-leg landing In addition, the knee joint indicated higher peak flexion moment (p<0.05) during double-leg landing than single-leg landing The hip and ankle joints illustrated greater elevated peak flexion moments (p<0.05) during single-leg landing than double-leg landing In the frontal plane, the hip showed larger peak adduction moments (p<0.05) than the knee and ankle joints during double-leg landing while the knee displayed larger abduction moments (p<0.05) than the hip and ankle joints during single-leg landing (Fig 5.1.15B) The hip also showed greater peak adduction moment (p<0.05) during double-leg landing than single-leg landing; the knee indicated larger peak abduction moment (p<0.05) during single-leg landing than double-leg landing

Fig 5.1.15: Comparison of peak hip, knee and ankle joint moments in the (A) sagittal and (B) frontal

planes during double-leg and single-leg landing

^ significant difference compared to single-leg landing (p<0.05)

# significance difference compared to double-leg landing (p<0.05)

* significant difference compared to knee joint (p<0.05)

** significant difference compared to other two joints (p<0.05)

84

e Joint power

In the sagittal plane, the knee exhibited greater joint power (p<0.05) than the ankle joint during double-leg landing; for single-leg landing, the ankle showed larger joint power relative to the hip and knee (Fig 5.1.16A) Furthermore, the knee has higher joint power (p<0.05) during double-leg landing than single-leg landing The hip and ankle possessed greater joint powers (p<0.05) during single-leg landing, compared to double-leg landing In the frontal plane, the hip showed larger joint power (p<0.05) relative to both knee and ankle while the knee has a higher joint power (p<0.05) than the ankle during double-leg landing (Fig 5.1.16B) During single-leg landing, the knee joint exhibited greater joint power (p<0.05) than both hip and ankle while the hip showed larger joint power than the ankle A higher joint power (p<0.05) was found at the knee joint during single-leg landing than double-leg landing

Fig 5.1.16: Comparison of peak hip, knee and ankle joint powers in the (A) sagittal and (B) frontal

planes during double-leg and single-leg landing

^ significant difference compared to single-leg landing (p<0.05)

# significance difference compared to double-leg landing (p<0.05)

* significant difference compared to knee joint (p<0.05)

+ significant difference compared to ankle joint (p<0.05)

** significant difference compared to other two joints (p<0.05)

f Eccentric work

In terms of eccentric work, the hip extensors and ankle plantarflexors delivered greater eccentric work (p<0.05) compared to the knee extensors during single-leg landing (Table 5.1.2) These muscles also exhibited larger eccentric work (p<0.05) during single-leg landing than double-leg landing On the other hand, the knee extensors showed higher eccentric work (p<0.05) during double-leg landing than single-leg landing For the frontal plane, the hip abductors and knee adductors displayed greater eccentric work (p<0.05), relative to the ankle adductors for both landing techniques The knee adductors also attained larger eccentric work (p<0.05) during single-leg landing than double-leg landing

Table 5.1.2: Summary of the mean eccentric work [standard deviation] and percentage contribution to

total energy dissipation of each lower extremity joint (hip, knee and ankle) in both sagittal and frontal planes during double-leg and single-leg landing

^ significance difference compared to single-leg landing (p<0.05)

# significance difference compared to double-leg landing (p<0.05)

* significance difference compared to knee joint (p<0.05)

+ significance difference compared to ankle joint (p<0.05)

** significance difference compared to hip and ankle joints (p<0.05)

Mean eccentric work (J/kg)

Contribution

to total energy dissipation (%)

Mean eccentric work (J/kg)

Contribution

to total energy dissipation (%)

Mean eccentric work (J/kg)

Contribution

to total energy dissipation (%)

Double-leg

-1.194 [0.393] 35.1

-1.200^

[0.303] 35.3

-1.009 [0.311] 29.7

-0.055+[0.081] 29.0

-0.008 [0.016] 4.3

-0.232**,#

[0.094] 60.7

-0.010 [0.010] 2.7

86

5.2 STAGE B – Investigation of Knee Injury Mechanisms

5.2.1 Porcine specimens during simulated landing impact

a Peak compressive force

Peak compressive force Fz ranged from 2112.6-4202.2N during the final compression trial; all specimens, except K3, underwent ACL failure (Fig 5.2.1A) Upon ACL failure, a significant drop in Fz response (1812.5-2659.3N) was observed, indicating an abrupt loss of loading in the joint structure (Fig 5.2.1B) Accumulation

of micro-damages in the ACL was clearly represented in the Fz drop profile; Fz drop estimated the maximum loss in loading through micro-damages sustained during compression The initial portion of the profile marked the steady rise in micro-damage prior to considerable ACL failure; the final portion, as indicated by the drastic increase, signified significant ACL failure The Fz drop in (+)ACL failure was significant compared to that in (-)ACL failure (p<0.01) (Table 5.2.1)

Fig 5.2.1A: Dissection photograph and magnetic resonance (MR) scans, pre-test and post-test, of a porcine knee specimen Black arrow indicates region of major ACL failure

Fig 5.2.1B: Profiles of the drop in compressive force Fz response during the successive displacement trials The gradual increase in F z drop at the initial portion indicated progressive accumulation of microdamages in the soft tissue, especially the ACL The drastic rise at the final portion represented significant ACL failure

Table 5.2.1: Experimental data, including failure type, of each specimen for joint compression during

(+)ACL failure Positive values indicate compressive force, posterior translation and external rotation; negative values indicate internal rotation F z drop (difference between peak F z during joint compression and mean F z during post-compression time period 300-500 ms) was compared for (-) and (+)ACL

failure [* indicates significance difference compared to (-)ACL failure (p<0.05)]

Note: K3 (shaded row) did not undergo ACL failure; data was derived from the final compression trial where cement support failure occurred

Within 10-ms time frame of

Change in Relative Posterior Femoral Displacement D x

(mm)

Change in Axial Tibial Angle

θ z (deg)

F z drop* F z drop

K1

ACL femoral

avulsion

K2

ACL femoral

avulsion

K4

ACL femoral

avulsion

K5

ACL tibial avulsion

K6

ACL femoral

88

b Anterior tibial translation and axial tibial rotation

During the 10-ms time frame of peak Fz (Fig 5.2.2), the specimens which had ACL failure exhibited a change in relative posterior femoral displacement Dx of 7.4-13.1mm and a change in axial tibial angle θz of –7.0-6.4 deg Conversely, K3 (which did not indicate ACL failure after dissection) presented changes in Dx and θz at 0.2

mm and –0.2deg respectively Essentially, K1 and K2 were in external rotating motion (increasing θz: 1.5 and 6.4deg) while K5 and K6 showed internal rotating motion (decreasing θz: -6.5 and –7.0deg) K4 did not display considerable change in

θz (-0.4deg) at peak Fz (Table 5.2.1)

Fig 5.2.2A: Specimen responses of a typical porcine specimen that attained ACL failure The

compressive force F z response indicated a peak F z during joint compression and a subsequent drop, followed by a post-compression plateau F z drop was estimated from the difference between the peak F z

and the mean F z during post-compression time period 300-500 ms The significant F z drop during (+)ACL failure, as compared to that in (-)ACL failure, presented a major ACL damage Within the 10-

ms time frame of the peak F z , the specimen responses in terms of axial tibial rotation θ z and relative posterior femoral displacement D x were noted These data provided information on the type of rotation (internal/external) and the presence of posterior femoral translation during peak F z

Fig 5.2.2B: Comparison of the drop in compressive force (Fz ) response between (-) and (+)ACL failure A substantial F z drop was observed for the specimens during ACL failure The changes in axial tibial angle θ z during peak F z indicated external rotation for K1 and K2, and internal rotation for K4 (very low), K5 and K6 The corresponding changes in relative posterior femoral displacement D x

responses showed considerable posterior translation upon peak F z

5.2.2 Human cadaveric specimens during simulated landing impact

a High-speed video capture of impact compression

Upon compressive impact loading, the tibia moved down and caused the femur to translate posteriorly; there was also a corresponding axial tibial rotation ACL failure was detected by a sudden jerk at the knee joint which indicated a loss of ACL restraint This occurrence of the sudden jerk was represented by a significant drop in compressive force response

90

Fig 5.2.3A: Key observations during impact loading in a typical ACL failure trial Images represented

selected frames of a high-speed video capture (sampling rate 1000Hz) Blue bar shows tibial motion while red bar shows femoral motion Yellow arrow signifies the direction of motion

Fig 5.2.3B: Responses of a typical human cadaveric knee specimen during pre-failure and ACL failure

The compressive force F z response described a peak compressive force F peak during impact loading and

a subsequent drop, followed by a post-compression equilibrium force F eqm F drop was estimated from the difference between F peak and mean F eqm during post-impact (300-500ms) The significant F drop upon ACL failure (as compared to pre-failure) demonstrated a major loss of anterior restraint on the tibia by the ACL The corresponding relative posterior femoral displacement D x and axial tibial rotation θ z

during the 10-ms time frame of F peak were also recorded for analysis on posterior femoral translation and type of rotation (internal/external) achieved by the specimen during F peak

b Compressive force, posterior femoral displacement and axial tibial rotation

All the specimens incurred ACL failure at actuator displacements between 22mm (Table 5.2.2A and Fig 5.2.4A) For each specimen, Fpeak was the highest (1.9-7.8kN) at the final compression trial in which ACL failure occurred During Fpeak, posterior femoral displacement (7.6-18.0mm) and internal tibial rotation (0.6-4.7deg) were observed (Table-5.2.2B) Additionally, there was a significant compressive force drop Fdrop of 79.8-90.9% (p<0.05) during ACL failure trials, compared to pre-failure trials (Fig.5.2.4B and C)

10-Table 5.2.2A: Summary of the impact compression trials conducted for each specimen Displacement

magnitude indicated the displacement value required to be achieved within 50 ms (based on a single 10-Hz haversine curve) within each trial Shaded boxes referred to the ACL failure trials

Table 5.2.2B: Experimental data of each specimen during ACL failure upon impact loading All

specimens underwent mid-substance ACL tear Positive values represent compressive force, posterior

translation; negative values represent internal rotation

ACL failure

10-ms time frame of F peak

Specimen

Force Fpeak (kN) Change in Posterior Femoral

92

Fig 5.2.4A: Pre-test and post-test magnetic resonance (MR) scans, and dissection photograph of a

human cadaveric knee specimen Black arrow shows major ACL failure

Fig 5.2.4B: Comparison of % Fdrop during pre-failure and ACL failure trials for all specimens

* significant difference compared to pre-failure (p<0.05)

Fig 5.2.4C: Gradual development of micro-damages resulted in weakening of the ACL structure This

weakening effect was illustrated by the steady increase in F drop with incremental displacement loading The drastic increases in F drop at the final displacement trials indicated a major structural failure in the ACL

5.2.3 Osteochondral damage in porcine specimens

a Histological observations

In general, the sites (A, E and P) displayed significant surface fraying and occasional clefts right down to the deep zone; on the contrary, I site did not present considerable surface irregularities but was observed with tidemark disruption in 5 out

of 6 specimens (Fig 5.2.5)

Fig 5.2.5: Photomicrographs (Hematoxylin & Eosin and Safranin-O/Fast Green staining) of

histological sections from control and post-impact specimens at the anterior (A), exterior (E), posterior (P) and interior (I) explant sites (Black bar = 200µm) In this figure, the presence of surface fraying and delamination (marked by bold arrow) was observed in sites A, E and P while tidemark disruption (marked by dotted circle) was noted in sites E and I

94

b Mankin score distribution

Modified Mankin grading of the histological sections for the post-impact specimens, K1 and K2, illustrated relatively higher scores at the anterolateral (AL/EL) and posteromedial (EM/PM) tibial cartilage compartments; on the other hand, K5 and K6 displayed relatively higher scores at the anteromedial (AM/EM) and posterolateral (EL/PL) compartments K4 showed higher scores at the posterior regions (PL/PM) of both compartments Specimens with ACL failure generally scored lower at the interior (IL/IM) sites; K3 (which did not have ACL failure) scored higher at the IL site The scores also revealed that joint compression caused tidemark disruption in most specimens (Fig 5.2.6)

Fig 5.2.6: Histological scoring based on the modified Mankin grading system The template displays

the 8 tibial cartilage sites where the explants were extracted A, E and P sites represent the anterior, exterior and posterior regions of the cartilage covered by the meniscus respectively while the I site is the exposed interior region; M and L refers to medial and lateral compartments respectively The Mankin score was stated at each explant site and an underlined score indicated tidemark disruption while a specimen label followed by an asterisk(*) represented ACL failure The control, which did not undergo joint compression, was scored to provide a baseline comparison For the post-impact specimens, grading was conducted and within each compartment, the sites were ranked according to their scores by the level of shading There are four levels of shading given to each site; the site with the lowest score within the compartment is unshaded while the site with highest score received the greatest shading level The distribution of shading specified the localization of the impact in each compartment

96

c Mankin score comparison between specimens

All specimens with ACL failure had significantly higher mean Mankin scores than the control (p<0.05); only K4, K5 and K6 had significantly greater mean Mankin scores than K3 which had no ACL failure (p<0.05) Generally, the A, E and P sites have significantly larger mean Mankin scores than the I sites (p<0.05) (Fig 5.2.7)

Fig 5.2.7: (A) Comparison of the mean overall modified Mankin scores for each specimen (with ACL

failure) with K3 (no ACL failure) and control [+ K4, K5 and K6 have mean scores significantly higher

than K3 (p<0.05) (K1 and K2: p=0.092 and 0.078 respectively); * All specimens except K3 have mean

scores significantly higher than the control (p<0.05)]

(B) Comparison of the mean modified Mankin scores at the anterior (A), exterior (E) and posterior (P) explant sites with the exposed interior (I) site [* A, E and P sites have mean scores significantly higher than the I site (p<0.05)]

5.2.4 Osteochondral damage in human cadaveric specimens

a Cartilage volume

In terms of cartilage volume, substantial reduction (p<0.05) was noted only at the lateral tibiofemoral compartments post-test (Fig.5.2.8)

Fig 5.2.8: Cartilage volumes of the medial and lateral tibial compartments and femoral condyles

obtained from segmentation of pre-test and post-test MR scans of each specimen Significant reduction

in cartilage volume was noted at the lateral tibiofemoral regions

* significant difference compared to pre-test (p<0.05)

b Cartilage thickness

There was also a significant reduction (p<0.05) in cartilage thickness post-test

at the anterior, posterior and interior cartilage regions for lateral tibial compartment, and the anterior, exterior and interior regions for medial tibial cartilage (Fig.5.2.9A) Marked thickness reduction (p<0.05) was also observed at the exterior, posterior and interior regions of lateral femoral condyle, and the anterior, posterior and interior regions of medial femoral condyle (Fig.5.2.9B)

98

Fig 5.2.9: Comparison of the (A) tibial and (B) femoral cartilage thickness pre- and post-test at the

anterior [A], exterior [E], posterior [P] and interior [I] regions of both medial and lateral compartments

* significant difference compared to pre-test (p<0.05)

100

c Histological observations

For the control specimens, mild-moderate surface irregularities were generally observed at the cartilage surface (Fig 5.2.10A) Damages noted in impacted specimens included surface fraying, delamination, mid-depth clefts and tidemark disruption; major osteochondral disruptions, especially at posterior tibiofemoral regions (Fig 5.2.10B)

Fig 5.2.10A: Photomicrographs (Hematoxylin & Eosin [H&E] and Safranin-O/Fast Green [Saf-O]

staining) showing the surface and the tidemark of histological sections from a non-impacted control specimen at the anterior [A], exterior [E], posterior [P] and interior [I] explant regions for both the tibial (left) and femoral (right) cartilages (Black bar = 200µm) The images generally depicted slight surface irregularities and moderate-extensive proteoglycan staining, but lacked the distinct damages present in impacted specimens

Fig 5.2.10B: Photomicrographs H&E and Saf-O staining) showing the surface and the tidemark of

histological sections from ACL-failed specimens at the anterior [A], exterior [E], posterior [P] and interior [I] explant regions for both the tibial (left) and femoral (right) cartilages (Black bar = 200µm) Bold arrow marks the presence of cartilage damage (surface fraying, delamination, clefts and tidemark disruption)

d Mankin score profiles

Mankin score profiles for all control specimens displayed mediocre scores across the tibiofemoral cartilages (Fig 5.2.11A) As for impacted specimens, high damage scores were generally more clustered around the anterior, exterior and

102

posterior tibial regions; for femoral cartilage, most damages were aggregated around the exterior, posterior and interior regions (Fig 5.2.11B)

Fig 5.2.11A: A, E, P and I regions refer to the anterior, exterior, posterior and interior regions of the

tibiofemoral cartilages respectively; at the tibial cartilage, A, E and P regions are covered by the menisci while I regions are exposed M and L represents the medial and lateral compartments respectively Mankin score profiles of the non-impacted control specimens indicated a distribution of mild-moderate scores throughout the tibial and femoral cartilages

Fig 5.2.11B: Cartilage damage profiles of the specimens based on the modified Mankin scoring

system The cartilage damage profile is described by the Mankin scores of each explant region Within each compartment, the regions were ranked according to their damage scores by the level of shading There are four levels of shading; the region with the least score is unshaded while the region with the largest score is allocated the highest shading level The shading distribution within each compartment determines the impact localization at the articular cartilage

104

e Mankin score comparison between regions

Histological results further showed that the region-specific Mankin scores for impacted specimens were significantly higher (p<0.05) compared to non-impacted control specimens, specifically at the tibial exterior medial, posterior medial and posterior lateral regions, and the femoral exterior, posterior and interior regions in both medial and lateral compartments (Fig.5.2.12A) Among the impacted specimens, there was no notable difference in Mankin scores (p>0.05) between regions for medial and lateral tibial cartilages; however, a substantial difference (p<0.05) was noted at the exterior and posterior regions of medial femoral cartilage, and the exterior, posterior and interior regions of lateral femoral cartilage, compared to the anterior regions (Fig 5.2.12B)

Fig 5.2.12A: Comparison of the region-specific Mankin scores at anterior [A], exterior [E], posterior

[P] and interior [I] regions for both medial [M] and lateral [L] compartments of the tibial cartilage

* significant difference compared to control (p<0.05)

Fig 5.2.12B: Comparison of the region-specific Mankin scores at anterior [A], exterior [E], posterior

[P] and interior [I] regions for both medial [M] and lateral [L] compartments of the femoral cartilage

* significant difference compared to control (p<0.05)

#

significant difference compared to AL region (p<0.05)

# #

significant difference compared to AL and AM regions (p<0.05)

5.2.5 Finite element analysis of a tibiofemoral joint during simulated landing impact

a Validation with experimental results

In comparison with the experimental results, the model displayed moderate similarity in terms of compressive force, posterior femoral displacement (relative anterior tibial translation) and axial tibial rotation at 6-mm actuator displacement (Fig 5.8.1)

106

Fig 5.2.13: Validation of model predictions, in terms of compressive force, posterior femoral

displacement and axial tibial rotation angle, with experimental results

b Prediction of compressive force, relative anterior tibial translation and axial tibial rotation

At 8-mm actuator displacement, the model predicted an increase in peak compressive force to 2.1kN, elevation in final posterior femoral displacement to 11.9mm and a drop in final axial tibial rotation angle to 2.0deg (Fig 5.2.14A)

Fig 5.2.14A: Model prediction of compressive force, posterior femoral displacement and axial tibial

rotation angle at 8-mm actuator displacement

In addition, at 10-mm actuator displacement, the model predicted a further increase in peak compressive force to 2.3kN, elevation in posterior femoral displacement to 11.9mm and a drop in axial tibial rotation to 1.1deg (Fig 5.2.14B)