Long Term Performance Testing

To ensure prolonged service, pre-clinical evaluation of an artificial joint is required to assess for wear damage. The present design involves a new concept of a bearing couple replacing the diseased joint and no joint simulator for testing and no suitable industrial standard for conducting such a test is available, so a new simulation machine had to be devised.

A finger joint motion simulator was fabricated for conducting wear and fatigue studies of the new artificial joint. The simulator can perform wear and fatigue tests on 3 artificial finger joints simultaneously under the same testing conditions. The motion generator simulates a planar motion from flexion-extension of a finger joint with one rotational degree of freedom. It is actuated by an electric motor with controllable speed to provide motion from 0° to 110c flexion. The proximal component of the finger joint is mounted on the upper rotating jaw for motion from flexion and extension while the lower jaw containing the distal component is connected to a pneumatic piston applying a physiologic joint load up to 500 N during each cycle. A load cell is installed in between the pneumatic piston and the artificial joint to monitor loading throughout the test. The complete setup is controlled by a programmable logic controller. Various conditions can be written to the system such as constant loading for a wear test and cyclic loading for fatigue. The position of the joint flexion and extension is monitored by means of a digital encoder to provide real time recording of the joint load with respect to the corresponding angular joint position.

The motion simulator (Fig. 9.12) enables cyclic loading and varied motion frequency to be applied to the test specimen, so that wear and fatigue characteristics can be assessed in a single run. Motion was set between 0° and 90° to simulate a normal range of flexion-extension in a PIP joint. The simulator employs a free movement cycle of the finger joint at a light loading between

10 and 15 N, running at 3 Hz. Following this a heavy pinch function cycle runs at 1 Hz applying a loading up to 110 N. The ratio between free movement and heavy pinch was set to 50:1, estimated from Joyce's study [75]. The simulator

Fig. 9.12 Joint motion simulator for wear testing

runs for 500 free movements then 10 heavy pinches and stops after half million cycles for component examination.

Six sets of the new design were subjected to simulator study. They were formed by precision EDM machining and grit blasted to produce a rough surface on the fixation stem. The bearing surface was polished by metallurgical means to a surface finish below 0.05 mm Ra, which is comparable to standards for commercial hip and knee implants. Using a coordinate measuring machine, the average clearances between all proximal and distal bearing couples after polishing were measured to be 0.56 mm. One pair of specimens was used as a control without being subjected to mechanical loading to study possible weight change from reacting with the simulated chemical environment alone. The other five pairs were mounted onto the fixture by clamping.

Each testing chamber was topped up with diluted bovine serum solution with 75% deionised water according to known standards for hip (ASTM F1714) and knee joints (ASTM F1715). The solution was supplemented by 0.2% sodium azide to retard bacterial growth and EDTA to prevent precipitated calcium phosphate on the articulating surface of the test specimen. The solution was filtered to remove particulates which could potentially result in third body wear. A quartz lighting system was used to warm the serum to about 37°C, which was then monitored and controlled using a thermocouple. New serum was added periodically to bring the level back to the original value, and to maintain the concentration as constant as possible. All tests were performed at a dual cycle mode to a total of 5 million cycles. The simulator was stopped every 0.5 million cycles to enable gravimetric measurement of the specimens using a microbalance with a resolution of 0.01 mg. The surface topography of the specimens was assessed quantitatively by non-contact 3-D interference microscopy for roughness (Ra) data, and qualitatively using scanning electron microscopy. The serum solution with suspended wear debris from each lubricating chamber was collected and stored at -18°C to allow isolation and characterization of the wear particles later on.

The serum lubricant was centrifuged at 2000 g for 60 min to produce a pellet of metal wear particles. The pellet was then suspended in 4 mL filtered distilled water with 3 mL 12 M potassium hydroxide (BDH) at 60°C for 48 h by shaking and sonicating frequently. Lipids and proteins were removed by extraction with 2:1 chloroform/methanol and repeated washes with 50% acetone. The metal wear particles were separated by sonication and filtered onto 0.1 mm polyester membranes. A section of each filter was analysed using a field-emission SEM at high magnification.

Energy dispersive X-ray analyses were performed to determine the elemental compositions of the wear particles. Captured images were analyzed using software to characterize the particle size distribution according to ASTM F1877.

The evaluation of the new PIP joint by the simulator study was stopped after 5 million flexion extension cycles. Wear damage of the distal component was most prominent with visible depressions due to material loss. Damage also extended to the central guiding track and the ridges of the distal and proximal components respectively but the overall profile was maintained. Magnification at 200x revealed fine linear scratches parallel to the joint motion but no crack or contact fatigue damage was observed.

The wear damage indicates average material loss of 14.57 mg and 10.99 mg of the proximal and distal components per half million cycles. The proximal component has a larger articulating surface area and longer sliding distance so wear loss was slightly higher than the distal component. Based on a density of 9.13 g/cm3 for CoCr and areas of the contact surfaces of 112.06 mm2 and 57.35 mm2 for the proximal and distal components, the cumulative wear volume and respective thickness loss can be calculated. The total volumetric wear and thickness loss after 5 million cycles were 15.95 mm3 and 142.37 mm for the proximal component, and 12.04 mm3 and 209.95 mm for the distal component. This corresponds to volumetric wear of 3.19 mm3 and 2.41 mm3, and thickness loss of 28.47 mm and 41.99 mm per million cycles, for the proximal and distal components respectively. Volumetric wear was found to be similar to MOM articulation in the hip joint (0.12-11.2 mm3/year). However, thickness loss of the finger joint model was relatively high, since standard hip joints observe thickness loss no greater than

10 mm/year [76]. This might be due to the small chamfers on the component edges of the finger joint design. The chamfer radius should therefore be enlarged in future designs to reduce material wear.

The surface topography was assessed using interference microscopy. Scanning was oriented perpendicular to the linear sliding scratches so that the maximum numbers of peaks and valleys could be taken into account. The (arithmetic) average roughness (Ra) was measured after each intermediate stop of half a million cycles. The specimen after fine polishing produced a mean Ra value below 0.03 mm. The surface roughness increased to a peak value of 1mm at the first half million cycles, as is typically observed in most simulator studies as a ''bedding-in'' phenomenon. A self-polishing action was observed subsequently, corresponding to a decrease of the surface roughness until reaching a plateau from 1.5 to 3 million cycles to 0.7 mm. There was then increase in surface roughness after 3 million cycles, showing accelerated wear and loss of the edge chamfer.

EDX measurements on the wear debris showed the same composition as the F90 CoCr alloy. The wear particles have an average value of 0.178 mm, and range from sub-micron to a few microns. These results are in good agreement with MOM articulations of the hip joint.

9.14 PIP Joint Stability

A well functioning finger joint comprises good range of motion and adequate stability to interact with demanding activities of the hand. Clinical instability of finger joints presents a major complication which significantly alters the hand function of those suffering from arthritic diseases. Contributions to a stable PIP

joint rely on integrity of three major elements of the bony geometric constraint, the musculotendinous pull and the ligamentous support. Ligamentous support is a primary stabilizer against external distraction forces and instantaneous joint loads from dynamic motion [77]. Under pathologic disturbances, which are predominantly rheumatoid arthritis and the corresponding inflammation processes, attenuation of the capsuloligamentous support commonly results in functional impairment of the hand from imbalanced kinematic chains and gross digital deformity [78]. This design was based on the rationale of a semi-constrained articulation, with supplementary stability to counterbalance the attenuated ligamentous support. Biomechanical evaluation of the lateral stability of the new design was performed to compare with the natural human joint and other commercial implant systems.

A kinematics test rig was built to assess the degree of lateral constraint in response to a radial distraction force. Different joint loads were applied to simulate an in-vivo condition by musculotendinous pull. The degree of lateral constraint was then assessed in terms of the magnitude of the resisting constraint force measured by a force sensor with a continuously applied angular lateral displacement to the articulation. A rotating table was constructed between the mounting fixtures of the proximal and middle phalanxes to allow the effects of a varying joint flexion angle to be studied. Mounting fixtures were specially designed to permit examination of the new design and other commercial implant systems, as well as natural phalangeal bones using the same testing conditions.

To minimize the effects ofthe large variations in mechanical properties ofthe soft-tissue structure, the setup adopted a simplified kinematic chain model from Stokoe's simulator [79]. This has been demonstrated to be compatible to the loading pattern of an intact joint and employs multiple mechanical pulleys and tension wires to mimic the configuration of a natural finger joint. Deadweight was added to tension wires to maintain a constant joint load across the articulation independent of the joint displacement. The test specimen was actuated by the loading axis of a testing machine, which provided continuous dislocation forces at quasi-static conditions (Fig. 9.13). The holder for the distal joint component was mounted on the moving axis of the machine to provide a dislocation motion against the proximal component. The holder for the proximal joint component was mounted onto the rotating table at a specific flexion angle and kept in stationary position on each test run. Since the angular displacement to dislocate the articulation from varied joint designs may not be linearly related to the motion of the machine axis, a magnetic resonance angular encoder with negligible frictional was added to the rotating axis of the middle phalanx so real-time angular displacement of the articulation could be accurately monitored.

Comparative studies of the lateral stabilities of the new PIP joint design and three commercial PIP joint implants from Ascension (Pyrocarbon PIP), Wright (Swanson PIP), and Depuy (Neuflex PIP) were made. The 7.5 mm proximal and distal joint components of the new joint design and the respective size of other

implants were mounted horizontally onto phalangeal implant holders. For the new design and the Pyrocarbon PIP implant, both using an unconstrained surface design, silicon oil was added between the joint surfaces to diminish frictional effects.

Testing was also conducted on 6 cadaveric middle-finger bone joints of similar sizes. The joint capsule and collateral ligament were released to isolate the effects from soft-tissue constraint to examine solely the geometric constraint from a bony articulation. Normal saline was added on the articular cartilage to keep the bone joint moist and serve as a lubricant. Examination of a cadaveric intact joint was also performed to reveal the compound effect from ligamentous and bony geometric constraints. This test, however, was only done in the full extension position.

All tests applied lateral displacements to the distal joint component up to 10c angulation from the neutral position, in accordance with the measured normal physiologic laxity of a PIP joint [80]. A constant load of 250 g was applied to the flexor tension wire, 200g to the extensor tension wire and 50g to each intrinsic tension wire producing a summation of joint load of 5.4 N across the articulation, based on the methodology of Uchiyama [81]. A series of constant joint loads of 16.2, 26.9, 37.8 and 48.6 N were applied on each test to account for an increasing musculotendinous pull across the articulation. This covers the normal range from isometric hand function [82]. Based on mean bone length of 26.21 mm, radially imposed loading applied by the cross-head was set perpendicular to the bone axis and kept constant at 25 mm distal from the PIP joint center, similar to the approach of Kiefhaber [80]. A 4N imposed loading would therefore represent a 0.1 Nm bending moment to the joint center according to this moment arm. The load was applied by displacement control at a rate of 0.5 mm/s [83] for 6 complete cycles to eliminate reading errors and dynamic effects of the joint components. Tests on five different constant joint loads and four varied flexion angles (0°, 30°, 60°, 90°) were performed to cover the physiologic range on each implant type and the bony joints. The reaction forces from resisting the angular displacement were monitored and recorded according to the degree of joint angulation.

Dynamic responses were observed to have restraining forces from the new PIP joint design consistently higher than the natural bony joint and all other commercial implant types. The constraining forces, obtained by increasing the joint load across the articulation, was highest for the new PIP joint design, followed by the natural bony joint, the pyrocarbon implant and lowest for the Swanson and Neuflex implants. Increasing the joint flexion angle generally has little effect on the constraining forces in the Swanson and Neuflex implants. The constraining forces produced in the Pyrocarbon implant were continuously raised by increasing the joint flexion angle from 0° to 90°. The constraining forces in the new PIP joint design were similar to the natural bony joint and found to be maximum at a flexion angle of 60° in agreement with the findings of Minamikawa [59].

The dynamic behavior falls into two types according to the mechanism of the implant design. The silastic type implants by Swanson and Neuflex showed a gentle increase in the constraining force without an apparent peak value with increasing lateral angular displacement up to the set point of 10°. In contrast, the unconstrained designs in the new PIP joint and the Pyrocarbon implant showed a bi-linear behavior with an initial step increase in the constraining force, followed by a slight reduction from the peak value. This bi-linear characteristic is comparable to a real bonyjoint presumably because of the anatomic compatibility of the surface design in these implants with the natural articulation. The stabilized surface geometry of the new design leads to an overall 30% increase in the maximum constraining force compared to natural bony joints, and a 68% increase compared to the Pyrocarbon implant. These results contrast with the joint stability from the silastic-type implants, which only exhibited 34% and 22% respectively of the constraining force of a natural bonyjoint.

The average values of the maximum constraining forces produced from each kind of articulation are compared with the intact human joint in Table 9.1. Without a joint load the intact joint maintains a stable position presumably because of the effective ligamentous constraint. Geometric constraints from the natural bony joint and all implant types without an effective joint load show a lack of joint stability. With increasing joint load up to the maximum physiologic

Table 9.1 Comparison of results to the constraint forces at different articulations at full extension joint position

Average constrain forces at different joint load conditions (N)

Table 9.1 Comparison of results to the constraint forces at different articulations at full extension joint position

Average constrain forces at different joint load conditions (N)

Test Specimen

5.4 N

16.2 N

26.9 N

37.8 N

48.6 N

Intact PIP joint

11.21 ±

11.80 ±

12.44 ±

12.80 ±

13.91 ±






Bony joint

1.47 ± 0.74

2.93 ± 1.47

4.75 ± 1.73

7.43 ± 1.63

10.87 ±


New PIP joint

























± indicates values of standard deviation

range, joint stability of the natural bony joint and the new joint approach the stability of an intact joint. Geometric constraints from the Pyrocarbon implant show some insufficiency when compared to the natural bone joint. Both the Swanson and the Neuflex implants indicated a lack of joint stability compared to the intact joint.

0 0

Post a comment