Faculty Bibliography

The anterior-posterior (AP) stability of the knee is an important aspect of functional performance. Studies have shown that the stability increases when compressive loads are applied, as indicated by reduced laxity, but the mechanism has not been fully explained. A test rig was designed which applied combinations of AP shear and compressive forces, and measured the AP displacements relative to the neutral position. Five knees were evaluated at compressive loads of 0, 250, 500, and 750N, with the knee at 15 degrees flexion. At each load, three cycles of shear force at +/-100N were applied. For the intact knee under load, the posterior tibial displacement was close to zero, due to the upward slope of the anterior medial tibial surface. The soft tissues were then resected in sequence to determine their role in AP laxity. After anterior cruciate ligament (ACL) resection, the anterior tibial displacement increased significantly even under load, highlighting its importance in stability. Meniscal resection further increased displacement but also the vertical displacement increased, implying the meniscus was providing a buffering effect. The PCL had no effect on any of the displacements under load. Plowing cartilage deformation and surface friction were negligible. This work highlighted the particular importance of the upward slope of the anterior medial tibial surface and the ACL to AP knee stability under load. The results are relevant to the design of total knees which reproduce anatomic knee stability behavior.

Horizontal cleavage tears (HCT) commonly occur in the posterior horn of the medial meniscus due to aging and degeneration. The purpose of this study was to investigate the surgical treatment of HCTs and their effect on dynamic tibiofemoral contact mechanics. The tibiofemoral contact mechanics of 10 cadaver knees were investigated using a custom dynamic loading apparatus, pressure sensor, and motion sensing camera. Three loading conditions were analyzed: 500 N compressive load, 500 N compressive load with 100 N posterior shear, and 500 N compressive load with 2.5 Nm of internal torque. Real-time peak contact pressures and contact areas were recorded throughout the full range of motion. After testing the intact meniscal state, a horizontal cleavage tear was created and included 50% of the width of the meniscus. The following procedures were performed, and the loading conditions described above were analyzed: HCT superior flap removal (5 specimens), HCT inferior flap removal (remaining 5 specimens), and both flaps removed (all 10 specimens). Statistical analysis was performed using a mixed linear effects model using the R-statistical package. The mixed linear effects statistical model identified statistically significant differences between independent variables, including the procedure performed, meniscal flap removed, meniscal region, loading condition, and knee flexion angle with respect to contact area and peak contact pressure. Peak contact pressure and contact area were not affected by selective flap removal (superior vs. inferior) or removal of both flaps of the HCT. We recommend that in the treatment of horizontal cleavage tears of the posterior horn of the medial meniscus, the outer 50% of the posterior horn of the medial meniscus should be maintained for load transmission.

The areas of the most frequent cartilage loss in mild-moderate medial osteoarthritis (OA) were reviewed from previous studies. Implant components were designed to resurface these areas. The surface geometries of the components were based on an average femur and tibia produced from 20 magnetic resonance imaging (MRI) models of normal knees. Accuracy of fit of the components was determined on these 20 individual knees. The femoral surface was toroidal, covering a band on the distal end of the femur, angled inward anteriorly. For a five-size system, the average deviations between the implant surfaces and the intact cartilage surfaces of 20 femurs were only 0.3 mm. For the tibia, the deviations were 0.5-0.7 mm, but the errors were mainly around the tibial spine, with smaller deviations in the central bearing region. Hence, these small implant components would accurately restore the original bearing surfaces and allow for preservation of all the knee structures. Using a thin metal component for the tibia would preserve the strong cancellous bone near the surface, an advantage for fixation. In this case, the femoral component would have a plastic bearing surface, but still be less than 10mm thickness. Such a design could have a useful place in the early treatment of medial OA of the knee.

A design concept was formulated for implants to treat medial osteoarthritis of the knee, using a metal plate resurfacing of the tibia plateau and a plastic bearing embedded in the distal end of the femur. A finite element analysis was carried out to determine whether a metal backing would be needed for the femoral component, and to what extent the stress and strain distribution in the trabecular bone surrounding the implant would match the normal intact condition. The CT scans from three knees scheduled for unicompartmental replacement were selected to generate computer models with variable bone densities in each element to cover a range of density patterns. Loading conditions were defined for a range of flexion angles, from loads at the center to the end of the component. A 2-peg fixation design was analyzed for both an all-plastic and a metal-backed construction. For the metal-backed, the interface von Mises stresses were close to intact values at the same level in the bone, although there was a 34 percent increase for loading at the end of the component. However, the all-plastic gave stresses elevated up to 109 percent. The maximum principal strain values for metal-backed in the trabecular bone below the implant were variable between specimens but close to intact under all conditions. In contrast the all-plastic showed strains up to 81 percent increased. The metal pegs showed load transfer, but the loads transmitted by the plastic pegs was small, as evidenced by the low interface stresses. The conclusion was that metal-backing was necessary to avoid excessive bone stresses and strains, while metal peg fixation was evidently an advantage.

BACKGROUND: Balancing is an important part of a total knee procedure, and in recent years, more emphasis has been given to quantifying the process. METHODS: During 101 total knee surgeries, initial bone cuts were made using navigation. Lateral and medial contact forces were determined throughout flexion using an instrumented tibial trial. Balancing was defined as a ratio of the medial and total force, the target being 0.5 (equal lateral and medial forces). Based on the initial values, surgical corrections were selected to achieve balancing. The most common corrections were soft tissue releases (63 incidences), including MCL, posterolateral corner, posteromedial corner, and changing tibial insert thicknesses (34 incidences). RESULTS: After final balancing, the mean ratio was 0.52 +/- 0.14, between 0.35 and 0.65 being achieved in 80% of cases. In 84% of cases, only 0-2 corrections were required. The average total force on the condyles was 215 +/- 86 N. CONCLUSION: Our study provides data to surgeons on the results to expect when balancing a knee, which can enhance both accuracy and consistency of the procedure.

BACKGROUND: The purpose of this study was to use MRI to determine if a loss of meniscal intra-substance integrity, as determined by T2* relaxation time, is associated with an increase of Kellgren-Lawrence (KL) grade, and if this was correlated with risk factors for cartilage degeneration, namely meniscal extrusion, contact area and anterior-posterior (AP) displacement. METHODS: Eleven symptomatic knees with a KL 2 to 4 and 11 control knees with a KL 0 to 1 were studied. A 3 Tesla MRI scanner was used to scan all knees at 15 degrees of flexion. With a 222N compression applied, a 3D SPACE sequence was obtained, followed by a spin echo 3D T2* mapping sequence. Next, an internal tibial torque of 5Nm was added and a second 3D SPACE sequence obtained. The MRI scans were post-processed to evaluate meniscal extrusion, contact area, AP displacement and T2* relaxation time. RESULTS: KL grade was correlated with T2* relaxation time for both the anterior medial meniscus (r=0.79, p<0.001) and the posterior lateral meniscus (r=0.55, p=0.009). In addition, T2* relaxation time was found to be correlated with risk factors for cartilage degeneration. The largest increases in meniscal extrusion and decreases in contact area were noted for those with meniscal tears (KL 3 to 4). All patients with KL 3 to 4 indicated evidence of meniscal tears. CONCLUSIONS: This suggests that a loss of meniscal integrity, in the form of intra-substance degeneration, is correlated with risk factors for cartilage degeneration.

BACKGROUND: Considerable debate remains over which anterior cruciate ligament (ACL) reconstruction technique can best restore knee stability. Traditionally, femoral tunnel drilling has been done through a previously drilled tibial tunnel; however, potential nonanatomic tunnel placement can produce a vertical graft, which although it would restore sagittal stability, it would not control rotational stability. To address this, some suggest that the femoral tunnel be created independently of the tibial tunnel through the use of an anteromedial (AM) portal, but whether this results in a more anatomic footprint or in stability comparable to that of the intact contralateral knee still remains controversial. QUESTIONS/PURPOSES: (1) Does the AM technique achieve footprints closer to anatomic than the transtibial (TT) technique? (2) Does the AM technique result in stability equivalent to that of the intact contralateral knee? (3) Are there differences in patient-reported outcomes between the two techniques? METHODS: Twenty male patients who underwent a bone-patellar tendon-bone autograft were recruited for this study, 10 in the TT group and 10 in the AM group. Patients in each group were randomly selected from four surgeons at our institution with both groups demonstrating similar demographics. The type of procedure chosen for each patient was based on the preferred technique of the surgeon. Some surgeons exclusively used the TT technique, whereas other surgeons specifically used the AM technique. Surgeons had no input on which patients were chosen to participate in this study. Mean postoperative time was 13 +/- 2.8 and 15 +/- 3.2 months for the TT and AM groups, respectively. Patients were identified retrospectively as having either the TT or AM Technique from our institutional database. At followup, clinical outcome scores were gathered as well as the footprint placement and knee stability assessed. To assess the footprint placement and knee stability, three-dimensional surface models of the femur, tibia, and ACL were created from MRI scans. The femoral and tibial footprints of the ACL reconstruction as compared with the intact contralateral ACL were determined. In addition, the AP displacement and rotational displacement of the femur were determined. Lastly, as a secondary measurement of stability, KT-1000 measurements were obtained at the followup visit. An a priori sample size calculation indicated that with 2n = 20 patients, we could detect a difference of 1 mm with 80% power at p < 0.05. A Welch two-sample t-test (p < 0.05) was performed to determine differences in the footprint measurements, AP displacement, rotational displacement, and KT-1000 measurements between the TT and AM groups. We further used the confidence interval approach with 90% confidence intervals on the pairwise mean group differences using a Games-Howell post hoc test to assess equivalence between the TT and AM groups for the previously mentioned measures. RESULTS: The AM and TT techniques were the same in terms of footprint except in the distal-proximal location of the femur. The TT for the femoral footprint (DP%D) was 9% +/- 6%, whereas the AM was -1% +/- 13% (p = 0.04). The TT technique resulted in a more proximal footprint and therefore a more vertical graft compared with intact ACL. The AP displacement and rotation between groups were the same and clinical outcomes did not demonstrate a difference. CONCLUSIONS: Although the AM portal drilling may place the femoral footprint in a more anatomic position, clinical stability and outcomes may be similar as long as attempts are made at creating an anatomic position of the graft. LEVEL OF EVIDENCE: Level III, therapeutic study.

Mechanical evaluation of total knees is frequently required for aspects such as wear, strength, kinematics, contact areas, and force transmission. In order to carry out such tests, we developed a crouching simulator, based on the Oxford-type machine, with novel features including a synthetic knee including ligaments. The instrumentation and data processing methods enabled the determination of contact area locations and interface forces and moments, for a full flexion-extension cycle. To demonstrate the use of the simulator, we carried out a comparison of two different total knee designs, cruciate retaining and substituting. The first part of the study describes the simulator design and the methodology for testing the knees without requiring cadaveric knee specimens. The degrees of freedom of the anatomic hip and ankle joints were reproduced. Flexion-extension was obtained by changing quadriceps length, while variable hamstring forces were applied using springs. The knee joint was represented by three-dimensional printed blocks on to which the total knee components were fixed. Pretensioned elastomeric bands of realistic stiffnesses passed through holes in the block at anatomical locations to represent ligaments. Motion capture of the knees during flexion, together with laser scanning and computer modeling, was used to reconstruct contact areas on the bearing surfaces. A method was also developed for measuring tibial component interface forces and moments as a comparative assessment of fixation. The method involved interposing Tekscan pads at locations on the interface. Overall, the crouching machine and the methodology could be used for many different mechanical measurements of total knee designs, adapted especially for comparative or parametric studies.

INTRODUCTION: Despite over four decades of development, there is still no established and universal method for the design, evaluation, and comparison of total knee replacement (TKR) designs. The intrinsic constraint of a TKR system is an important metric for determining the stability of a TKR, and there are standardized techniques for measuring these properties (e.g. ASTM F 1223-14). These standards describe the resistance to anterior-posterior (AP) and medial-lateral (ML) displacements, and internal-external (IE) rotation under an axial load between the femoral and tibial components at discrete flexion angles. Such tests, however, do not necessarily map directly to any physiologically relevant loading scenarios; thus, comparisons of TKR systems based on these techniques may provide misleading results. Recent advancements in the development of force-controlled joint motion simulators have allowed more physiological in-vitro joint loading, including the contributions of virtual soft tissues. We hypothesize that a function-based technique for measuring TKR constraint may yield a different classification of TKR systems in terms of constraint, as compare with the long-standing discrete-angle measurement technique. Therefore, the objective of this study is to develop an alternative, function-based means of assessing TKR constraint. METHODS: The intrinsic constraints of three commercially available TKR systems were measured by performing a series of tests. The systems considered spanned a range of congruencies between the femoral and tibial condyles. Design A was a relatively low-conforming cruciate-retaining design. Design B was a moderately-conforming cruciate-retaining design. Design C was ultra-congruent with asymmetrical medial versus lateral condyles, which can be used in cruciate-sacrificing TKR. For each specimen, the femoral component was affixed onto a custom fixture on the upper actuator of the VIVO using Bosworth Fastray (Keystone Industries, Gibbstown, NJ) polymethyl methacrylate (PMMA), such that the flexion axis of the femoral component coincided with the flexion axis of the VIVO. The tibial component was affixed into a fixture on the lower actuator in a neutral orientation with respect to the femoral component using dental stone. Flexion/extension and abduction/adduction motions were provided by the upper actuator, while all joint displacements and internal/external rotation were provided by the lower actuator. Prior to any force-controlled testing, displacements were applied in the AP, ML, and IE directions with an axial load of 710 N, in order to determine the maximum forces and moments each TKR system could safely withstand without dislocating or achieving clinically unrealistic displacements. Discrete Flexion Tests: Constraint tests were then conducted at discrete flexion angles (0degree, 30degree, 60degree, and 90degree), with the same axial load of 710 N. With all other DOFs unconstrained, a sinusoidal force or torque (period = 20 s), was applied in the AP, ML, and IE directions in succession. The maximum and minimum values of each sinusoidal load was equal to 80% of the previously determined limits. The motions in all DOFs were recorded by the joint simulator. This protocol essentially reproduced the ASTM standard constraint tests. Continuous Flexion Tests: Next, constraint tests were conducted by applying a continuous flexion/extension motion between 0degree and 90degree, with 710 N axial load, and all other DOFs unconstrained. The relative motions of the femoral and tibial components in the AP, ML and IE directions were recorded and defined the neutral path of motion of the TKR system. The same loading was repeated, but with a constant anterior, posterior, medial, lateral, internal, or external load applied, equal to 80% of the limit for that TKR system. The relative change in joint motion compared to the neutral path was measured for each superimposed load; this change was defined as the laxity. Physiological Motion Tests: Knee joint loads and flexion angles during gait have been previously reported in the literature, and were used as inputs for the VIVO joint simulator. The loads were scaled by a constant factor such that the maximum axial load was equal to 710 N to prevent any implant damage. The neutral path of each TKR during gait was determined throughout the simulated gait cycle. Then the laxity was measured by superimposing AP, ML, and IE loads as before, on top of the AP, ML, and IE loads already applied during the simulated gait cycle. The magnitudes of the superimposed loads were adjusted proportionately with the time-varying axial force in order to prevent large distracting loads from causing excessive displacements when there is little joint compression. Additionally, the cruciate and collateral ligaments were simulated as tension-only point-to-point springs incorporated into the control software of the VIVO joint motion simulator. Stiffness characteristics and origin / insertion points were obtained from the MB Knee: Multibody Models of the Human Knee project (simtk.org/home/mb-knee). The continuous flexion and physiological motion tests were performed with all four major ligaments disabled, all enabled, all enabled with each ligament individually disabled, and the collateral ligaments enabled with cruciate ligaments disabled. RESULTS: Figure 1 compares the AP laxity data generated using discrete flexion tests with the corresponding data from the continuous flexion tests, for all three TKR systems. Figure 2 demonstrates the effect of including simulated soft tissues on the resulting AP laxity during continuous flexion for all three TKR with two ligament configurations. The first configuration includes the PCL, LCL, and MCL; the second configuration has all ligaments disabled. Figure 3 shows the AP and IE laxities of the moderately-conforming cruciate-retaining TKR as measured using the physiological motion test, comparing two ligament model configurations (enabled and disabled). The flexion angle of the femoral component to the tibial component is also provided for reference. DISCUSSION: The AP laxities of the continuous flexion tests follow the same trends as the discrete tests, but higher absolute values are achieved at each flexion angle. These differences are likely related to differences in friction and simulator dynamics between the different tests. The continuous flexion tests demonstrated that when posterior loads are applied to the tibia, larger posterior displacements occur when the PCL is omitted, at 0degree to 55degree flexion. Since the ligaments provide restraining force to posterior tibial displacement, one would expect the absence of the ligaments would result in higher AP Laxity when a posterior load is imposed. Figure 2 confirms that the expected result occurs when the ligaments are disabled in the VIVO controls, validating the model. The final figure illustrates the importance of the inclusion of soft tissues in measuring TKR constraint. The AP and IE laxities of the TKR system under a functional loading profile vary substantially when ligaments are considered. SIGNIFICANCE: Current knee constraint tests do not correspond to physiologically relevant functional scenarios, and thus it may be misleading to extrapolate results from those tests to in vivo function. An alternative, function-based means of assessing TKR constraint could lead to advancements in design, as well as aid in prescribing suitable designs for current clinical use

INTRODUCTION: Implant loosening is an important failure mode in unicompartmental knee arthroplasty (UKA). Specifically, tibial implant looseningmay be caused in part by the 6-8 mm of bone resected during surgery. Removal of too much bone can result in inadequate support as studies have shown that strength and density of trabecular bone decreases rapidly with depth from the proximal tibia surface [1][2]. A main purpose in this study was to evaluate sinkage between different implant designs and their respective resection depths The first part of the study describes the methodology for testing the implants without requiring cadaveric specimens. Our group modeled and 3D printed a bone substitute with a realistic structure and density distribution which can be used in a laboratory setting for comparative experimental testing (Figure 1). 3D printing of bone eliminates some of the major challenges found when using human and animal specimens, or homogenous porous foams that do not accurately characterize natural bone. Furthermore, 3D printing bone provides researchers with an affordable alternative to perform uniform, comparative experiments. The second part of the study was to develop a testing method which would evaluate the fixation and sinkage of tibial implants. By incorporating the 3D printed trabecular bone, comparative tests were carried out on different unicompartmental knee replacement (UKR) designs. METHODS: A microCT scan of trabecular bone was obtained from the medial compartment of a healthy human proximal tibial allograft (SkyScanll72, Brucker MicroCT, Kontich, Belgium). The scanned image was segmented in Mimics 16.0 software (Materialise, Leuven, Belgium). A3D computer model was generated in Mimics and exported to Geomagic Design X (3D Systems, Rock Hill, SC) for further processing. We used a high resolution stereolithography machine (ProJet 7000HD, 3D Systems, Rock Hill, SC) to prototype our test samples in a resin material. Cylindrical samples of 8mm diameter and 1:1 scale and 15 mm overall lengths were fabricated andtested, and additional samples of 8 mm diameter and 19 mm length scaled up 25% were also printed andtested [3]. Mechanical load testing was conducted in a servo hydraulic load frame (MTS, MTS, Eden Prairie, MN) using a similar setup as Morgan and Keaveny 2001 [1], where compressive force was applied at a constant strain rate of 0.5% per sec until failure after three preconditioning cycles to 0.1% strain. From the original CT scan of tibial bone, a larger block was created through patterning the structure, and implant models were used to virtually cut the cavities for several implant designs. Once the bone samples were 3D printed, four different UKR designs were cemented into them (Figure 2). The Early Intervention (EI) component had a bone resection depth of 2 mm. The all plastic inlay component had a bone resection depth of 4mm. Amidlay design with aresection depth of 4 mm and an onlay design with aresection depth of 6 mm were alsotested. The samples werethen placed in a coordinate measuring machine (CMM) (Microgage, Hetnel, Niagara Falls, NY) to obtain nominal values. Samples were then loaded into a fatigue machine and testing ran for 100,000 cycles at 1Hz. After each 20,000 cycles, the samples were removed and again placed in the CMM and values were compared to the nominal values obtained at the beginning of the test. Measurements obtained from the CMM where then used to reconstruct implant placement in Geomagic. Models of the implants were aligned and subsidence was determined by mesh deviations (Figure 3). RESULTS: The modulus of the trabecular bone plugs was calculated from the cross sectional area and the force-displacement curve. For the 1:1 ratio plugs, the modulus was 27.9 MPa(n =6, a = 5.5 MP a), and the modulus ofthe scaled up 25% ratio plugs was 48.7 MPa (n =6, a= 11.8MPa). The cross sectional area of each specimen was 50.3 mm². Comparing to the modulus of tibial trabecular boneplugs ofthe same size, the results matched the low end of the range measured and the plugs failed in the region where the density was transitioning from dense subchondral bone to sparser trabecular bone. Results from the comparative sinkage evaluation for the four implant designs are shown in Figure 3. The EI type component demonstrated the lowest sinkage ofthe group over 100k cycles, and the all-plastic inlay component showed the highest sinkage. The results mimicked expected results predicted by finite element analysis [2]. DISCUSSION: Mechanical load testing ofthe 3D printed trabecular bone determined the characteristics matched well with those found in human samples and was a feasible substitute for testing. Observation ofthe tests also showed that the bone substrate fails in a similar mode and similar zone where the density transitions to less dense inner trabecular regions. Using the 3D printed bone allowed for uniformity in the testing samples giving better comparative results. The CMM measurements after each iterative cycle were compared to measurements taken before fatigue testing. Though more cycles would show long term effects, most looseninghappens within a much shorter period of time. This study showed that UKRs requiring less bone resection subsided less, suggesting better fixation. Implementing 3 D printed trabecular bone as a mechanical testing substrate for sinkage analysis can provide valuable information for UKR design and performance. SIGNIFICANCE: Synthetic 3D printed trabecular bone offers many significant advantages over human or animal tissue when used for mechanical testing. This technique eliminates inter-specimen variability of human or animal tissue, is less difficult to maintain in long term testing, will not degrade due to biological decay, has a geometric structure that matches the anatomic trabecular bone, simulates the density gradient in bone which other common testing materials cannot, andean be created in any shape from any bone or bone segment