Methods

Alignment specific physiological loading was determined using telemetric tibial implant data from Orthoload [2] and applying it to a validated finite element lower limb model developed by the University of Denver [3]. Two high demand activities were selected for the loading section of this study: step down (SD) and deep knee bend (DKB). Using the lower limb model, hip and ankle external boundary conditions were applied to the ATTUNE^® knee system for both MA and AA techniques. The 6 degree of freedom kinetics and kinematics for each activity were then extracted from the model for each alignment type. Mechanical alignment (MA) was considered to be neutral alignment (0° Hip Knee Ankle Angle (HKA), 0° Joint Line (JL)) and AA was chosen to be 3° varus HKA, 5° JL. It is important not to exceed the limits of safety when using AA as such it is noted that DePuy Synthes recommends staying within 3º varus HKA and 3º JL. The use of 5º JL was used in this study to account for surgical variation [Depuy-Synthes surgical technique DSUS/JRC/0617/2179].

Following a similar method described by Maag et al [1] ATTUNE tibial implants were cemented into a bone analog with 2 mL of bone marrow in the distal cavity and an additional reservoir of lipid adjacent to the posterior edge of the implant. Tibial implant constructs were then subjected to intra-operative ROM/stability evaluation, followed by a hyperextension activity until 15 minutes of cement curing time, and finally 3 additional ROM/stability evaluations were performed using an AMTI VIVO simulator. The alignment specific loading parameters were then applied to the tibial implants using an AMTI VIVO simulator. Each sample was subjected to 50,000 DKB cycles and 120,000 SD cycles at 0.8 Hz in series; approximating 2 years of physiological activity. After physiological loading the samples were tested for fixation failure load by axial pull off.

METHODS

Two sets of three ATTUNE^® size 5 right leg CR RP TKRs (DePuy Synthes, Warsaw, IN) were used in a study to evaluate ADL implant wear. Implant specific loading profiles were produced via a validated finite element lower limb model [1] that uses activity data such as gait (K1L_110108_1_86p), SD (K1L_240309_2_144p), and DKB (K9P_2239_0_9_I1) from the Orthoload database [2] to produce external boundary conditions. Each set of components were tested using a VIVO joint simulator (AMTI, Watertown, MA, Figure 1) for a total of 4.5 million cycles (Mcyc). All cycles were conducted at 0.8Hz in force-control with flexion driven in displacement control. Bovine calf serum lubricant was prepared to a total protein concentration of 18g/L and maintained at 37°±2°C. Wear of the tibial inserts was quantified via gravimetric methods per ISO14243–2:2009(E). Polyethylene tibial insert weights were taken prior to testing and every 0.5Mcyc there after which corresponded to serum exchange intervals. The multi-activity test intervals were split into10 loops of 1,250 DKB, 3,000 SD, and 45,750 gait cycles in series. Based on activity data presented by Wimmer et al. the number of cycles per activity and activities used is sufficient for a person that is considered active [3]. A loaded soak control was used to compensate for fluid absorption in wear rate calculations. Wear rates were calculated using linear regression.

Method

A method for creating rim fracture in vitro was developed to evaluate implant design features and polyethylene liner materials. Liners were secured into acetabular shells, fixtured in resin mounted at a 55° (in vitro; 65° in vivo) inclination to ensure high load/stress was at the area of interest. Ranges of kinematic and maximum applied load profiles were investigated (parameters summarized in Table 1). Testing was conducted on an AMTI 12-station hip simulator for 0.25–1.0 million cycles or until fracture (lubrication maintained with lithium grease). At completion, liners were cleaned and examined for crack propagation/fracture. Inspection of the impingement site on the opposite rim was also analyzed. Additional assessments included liner disassociation/rock out, deformation of characteristics such as anti-rotation devices and microscopic inspection of high-stress regions.

Materials and Methods

Previously reported DKB profiles were developed based on in vivo patellofemoral data and include a wide range of patient variability. The profiles chosen for this body of work were based on the stress in the patellar lateral facet; maximizing this stress whilst maintaining the ability to run the profile stably on the simulator. Load/kinematic profiles were run on three patellar designs (n=3 per group) for 220,000 cycles at 0.8Hz on an AMTI VIVO joint simulator. A comparison cohort of clinically retrieved devices of the same design was identified in an IRB-approved database. Exclusion criteria included gross delamination, cracking secondary to oxidation, and surgeon-reported evidence of malalignment leading to mal-tracking. 29 Patellae were included for analysis: PFC^® All Poly (n=14), ATTUNE^® Anatomic (n=6), and ATTUNE^®Medialized Dome (n=9). Mean in vivo duration was 70.1 months. Patellae were analyzed under optical microscope in large-depth-of-field mode to map the surface damage profile. Burnishing ‘heat-maps’ were generated for retrievals and simulated patellae by normalizing the patellar size and overlaying silhouettes from each component of the same type using a custom-developed MatLAB code.

METHODS

External hip and ankle boundary conditions were created for a previously-developed finite element lower limb model [1] for step down (SD), deep knee bend (DKB) and gait to best reproduce in-vivo loading conditions as measured on patients with the Innex knee (orthoload.com) (Figure1). These boundary conditions were subsequently used as inputs for the model with a current fixed-bearing total knee replacement to estimate implant-specific kinetics and kinematics during activities of daily living. Implant alignments were varied, including variation of the hip-knee-ankle angle-±3°, the frontal plane joint line −7° to +5°, internal-external femoral rotation ±3°, and the tibial posterior slope 5° and 0°. Through varying these parameters a total of 2464 simulations were completed.

A NN was created utilizing the NN toolbox in MATLAB. Sequence data inputs were produced from the alignment and the external boundary conditions for each activity cycle. Sequence outputs for the model were the 6 degree of freedom kinetics and kinematics, totaling 12 outputs. All data was normalized across the entire data set. Ten percent of the simulation runs were removed at random from the training set to be used for validation, leaving 2220 simulations for training and 244 for validation. A nine-layer bi-long short-term memory (LSTM) NN was created to take advantage of bi-LSTM layers ability to learn from past and future data. Training on the network was undertaken using an RMSprop solver until the root mean square error (RMSE) stopped reducing. Evaluation of NN quality was determined by the RMSE of the validation set.

Methods

Two polyethylene acetabular liner material test groups were chosen for this study: commercially available Marathon^® (A) and AltrX^® (B). All liners were machined to current production specifications with an inner diameter of 36mm and an outer diameter of 56mm. Surface roughness (Ra) of the liner inner diameters were measured using contact profilometry in the head-liner contact area, before and after 3Mcyc of wear testing. Liners were soaked in bovine serum for 48 hours prior to testing. Friction factor measurements were taken per ASTM WK28778 prior to, and after wear testing using an external six degrees of freedom load cell (ATI Industrial Automation) and a reduced maximum vertical load of 1900N.

Friction factor and wear testing was conducted in bovine serum (18mg/mL total protein concentration) supplemented with 0.056% sodium azide (preservative) and 5.56mM EDTA (calcium stabilizer) on a 12-station AMTI (Watertown, MA) ADL hip simulator with load soak controls per ISO 14242-1:2014(E). The liners were removed from the machine, cleaned and gravimetric wear determined per ISO 14242-2:2000(E) every 0.5 million cycles (MCyc) through a total of 3Mcyc to evaluate wear.

Methods

Three commercially available fixed bearing tibial implant designs were investigated in this study: ATTUNE^®, PFC SIGMA^® CoCr, ATTUNE^® S+. The implant designs were first prepared using a LMI implantation process. Following the method described by Maag et al tibial implants were cemented in a bone analog with 2 mL of bone marrow in the distal cavity and an additional reservoir of lipid adjacent to the posterior edge of the implant. The samples were subjected to intra- operative range of motion (ROM)/stability evaluation using an AMTI VIVO simulator, then a hyperextension activity until 15 minutes of cement cure time, and finally 3 additional ROM/stability evaluations were performed.

Implant specific physiological loading was determined using telemetric tibial implant data from Orthoload and applying it to a validated FE lower limb model developed by the University of Denver. Two high demand activities were selected for the loading section of this study: step down (SD) and deep knee bend (DKB). Using the above model, 6 degree of freedom kinetics and kinematics for each activity was determined for each posterior stabilized implant design.

Prior to loading, the 3-D motion between tibial implant and bone analog (micromotion) was measured using an ARAMIS Digital Image Correlation (DIC) system. Measurement was taken during the simulated DKB at 0.25Hz using an AMTI VIVO simulator while the DIC system captured images at a frame rate of 10Hz. The GOM software calculated the distance between reference point markers applied to the posterior implant and foam bone. A Matlab program calculated maximum micromotion within each DKB cycle and averaged that value across five cycles.

The implant specific loading parameters were then applied to the three tibial implant designs. Using an AMTI VIVO simulator each sample was subjected to 50,000 DKB and 120,000 SD cycles at 0.8Hz in series; equating to approximately 2 years of physiological activity. Following loading, micromotion was measured using the same method as above.

METHODS

GUR 1050 UHMWPE acetabular inserts, re-melted and cross-linked at 5.0Mrad (Marathon®, DePuy Synthes Joint Reconstruction, Warsaw, IN), were artificially aged per ASTM F-2003 in a stainless steel chamber at 5 atm. oxygen pressure and 70°C. Samples were maintained at temperature for 9, 10.4 and 11 weeks. After aging was completed, Fourier Transform Infra-Red (FTIR) spectroscopy was employed on one insert from each time point to evaluate the induced oxidation as a result of artificial aging. Resulting induced BOI values measured by FTIR were 0.195, 0.528 and 1.184. UHMWPE inserts had an inner diameter of 28mm and an outer diameter of 48mm and were articulated against 28mm diameter M-Spec® metal femoral heads (DePuy Synthes Joint Reconstruction, Warsaw, IN). Testing was conducted on a 12-station AMTI ADL hip simulator (AMTI, Watertown, MA) with load soak controls per ISO 14242-1:2014(E) in bovine serum (18mg/mL total protein concentration) supplemented with 0.056% sodium azide (preservative) and 5.56mM EDTA (calcium stabilizer). The UHMWPE inserts were removed from the machine, cleaned, and gravimetric wear determined per ISO 14242-2:2000(E) every 0.5 million cycles (MCyc) for 4.0 MCyc total. A two-tailed student's t-test was used (variance determined by F-test results) to analyze differences in wear rates between the three test groups.