METHODS

Fresh-frozen porcine acetabulae (n = 20) were prepared and embedded. Hemispherical reamers were used and the last reaming step was performed using a vertical drilling machine to ensure a proper alignment of the cavity axis. A hand-guided 3D laser scanner was used (HandySCAN 700, Creaform) to determine the reamer geometry and the cavity shape. Press-fit cups with two different surface coatings (Ø44 mm, Porocoat/Gription, DePuy Synthes) were implanted using a drop tower. The Porocoat cup was implanted with impacts from lower drop heights (low implantation force) and press-fits of 1 mm and 2 mm. The Gription cup, exhibiting a rougher surface, was implanted with low and high implantation forces and a press-fit of 1 mm. Bone-implant contact was analysed by the registration of the cup and cavity surface models, scanned prior to implantation, to the scan of the implanted cup. The cup surface was divided in areas with and without contact to the surrounding cavity. Overhang indicates that there was no adjacent cavity surface surrounding the implanted cup. The transition between contact and a gap at the cup dome was defined as contact depth and used as indicator for the cup seating.

Methods

30 right human femora were collected, fresh frozen and divided in six groups, matching for age, gender, height, weight and bone density. In group 1–3 a bone defect of AORI type F2a (15mm medial) and in group 4–6 a defect of AORI type F3 (25mm on both sides) was created. In all six groups the same modular femoral surface component (Endo-Model-W, Waldemar Link) was used, combined with different stem types (100/ 160 mm cemented / uncemented / standard/ anatomical with / without cone). Additionally, one trial was set up, omitting the modular stem. The correct fit of the implants was confirmed by fluoroscopy. After embedding, specimens were mechanically loaded 10mm medially and parallel to the mechanical femoral axis with an axial force of 2700N and a torsional moment of 5.6Nm at a flexion angle of 15° with respect to the coupled tibial plateau according to in-vivo gait load for 10,000 cycles (1Hz) in a servohydraulic testing machine (Bionix, MTS). The relative movement between implant, cement and distal femur was recorded using a stereo video system (Aramis3D,gom). An axial pull-out test at 1mm/min was performed after dynamic loading.

Material and Methods

16 titanium rods (Ti-6Al-4V, extra low interstitial elements, according to DIN 17851, ⊘ 5 mm, 120 mm length) were stress-relief annealed (normal atmosphere, holding temperature 622 °C, holding time 2 h) and cooled in air. An implant specific surface roughness was achieved by chemical and electrolytic polishing (Ra = 0.307, Rz = 1.910). Dry (n = 6) and wet (n = 6, 5 µl phosphate buffered saline) flashovers were applied with a hand-held electrode of a high-frequency generator (Aesculap AG, GN 640, monopolar cut mode, output power 300 W, modelled patient resistance 500 Ω). The size of the generated discoloured area on the rod's surface - representative for the heat affected zone (HAZ) - was determined using laser microscopy (VK-150x, Keyence, Japan). Rods without flashover (n = 4) served as control. The fatigue strength of the rods was determined under dynamic (10 Hz, load ratio R = 0.1), force-controlled four-point bending (FGB Steinbach GmbH, Germany) with swelling load (numerical bending stress 852 MPa with a bending moment of 17.8 Nm) until failure of the rods. The applied bending stress was estimated using a finite-element-model of a hip stem during stumbling. Metallurgical cuts were made to analyse the microstructure.

METHODS

The surfaces of two cups (ASR, DePuy, Leeds; one pristine, one a worn retrieval) were digitized using a coordinate measurement machine (CRYSTA-Apex S574, Mitutoyo; 3 µm accuracy). Both cups were measured undeformed and while being deformed between a clamp. Three different methods for generating reference geometries were investigated (PolyWorks|Inspector 2018, InnovMetric). Method 1: A sphere with the nominal internal cup dimensions was generated. Method 2: A sphere was fitted to the measured data points after removing those from worn areas (deviation > 3 µm is defined as wear) to eliminate the influence of manufacturing tolerances on the nominal diameter. Method 3: Measurements, which displayed visual deformation in the computed wear pattern based on the best fit sphere, were fitted with an ellipsoid. The direction of the deformation axes and the amount of deformation were used to scale the best fit ellipsoid. Linear wear was calculated from the distance of the respective reference geometry to the measured point cloud. Finally, material loss is defined as the difference in volume of the reference geometry and the measured geometry.

Material & Methods

The micromotions within two taper articulations were investigated: a head-stem taper (Corail, DePuy Synthes, Leeds, UK, Figure 1) and a neck-stem taper of a bi-modular THA prosthesis (Rejuvenate, Stryker, Kalamazoo, MI, USA). Both tapers were assembled with 2000 N. Loading at an angle of 50° to the taper axes (identical for both) in direction of the stem axis was incrementally increased from 0 N to 1900 N (n=3). Small windows (< 2.5 mm²) were cut through the female tapers by electric discharge machining, exposing the male taper surface for direct micromotion measurements by microscopic topographic measurements (Infinite Focus Microscope, Alicona Imaging GmbH, Austria). Subsequently, feature matching of the images from the differently loaded implants was applied (Matlab 2016b, The MathWorks Inc., Natick, MA, USA) to determine the local relative motion between the mating surfaces.

METHODS

THAs were performed by an experienced orthopaedic surgeon in full cadaveric models simulating real patient surgery (n=7). Acetabular cups with a Porocoat™ (n=3) and Gription™ surface coating (n=4) were implanted (DePuy Synthes, Leeds, UK). Computed tomography (CT) scans prior to surgery, as well as after reaming and implantation of press-fit cups were used to calculate the HRC displacement. After aligning the pelves in the anterior pelvic plane, 3D reconstruction of the HRC at each stage was performed by fitting spheres to the femoral head, the reamed cavity and the inserted cup. 3D surface models of the cups were generated using a laser scanner and were registered to the CT images. The effective press-fit was calculated using the diameters of spheres, fitted to the cavity prior to cup insertion and to the outer cup coating. The polar gap was defined as the difference between the outer cup surface and the subchondral bone at the cup pole. Anteversion and abduction angles were calculated as difference between the cup planes and the sagittal and transverse plane, respectively.

METHODS

Paired human femora (m/f=11/12, age=60±18 y) were scanned with quantitative computed tomography (QCT, Philips Brilliance 16) before broaching, with the final broach, after its removal and after stem implantation. Compaction broaching (n=4) was compared in an in situ (cadaver) study against extraction broaching with blunt tooth types (n=3); in an ex situ (excised femora) study, compaction broaching was compared against extraction broaching with sharp tooth types (n=8 each). QCT data were resampled to voxel sizes of 1×1×1 mm (in situ) and 0.5×0.5×1 mm (ex situ). Mean trabecular BMD of the proximal femur was determined. The cavity volumes were segmented in the post-broach images (threshold: −250 mgHA/cm3, Avizo 9.2) and a volume of interest (VOI) of one-voxel thickness was added around the cavity to capture the interfacial bone. VOIs were transferred to the pre-broach image and bone densification was calculated within each VOI as the increase from pre- to post-broach image (MATLAB). Detailed surface data sets of broaches and stems were collected with a 3D laser-scanner (Creaform Handyscan 700) and aligned with the segmented components in the CT scans (Fig. 1). Stem seating was defined as the difference between the top edge of the stem coating and the final broach. Distance maps between the stem and cavity surface were generated to determine the bone-implant contact area and press-fit. All parameters were analysed between 5 mm distal to the coating and 1 cm distal to the lesser trochanter and analysed with related-samples Wilcoxon signed rank and Spearman's correlation tests (IBM SPSS Statistics 22).

Materials & Methods

The ball heads (ø 32mm, 12/14, size L, CoCr or Al2O3) were quasi-statically assembled to the stems (Ti or CoCr, Metha, Aesculap AG, Germany, v=0.5 kN/s, F=6 kN, n=3 each, 10° adduction/ 9° flexion according to ISO 7206-4) and then loaded sinusoidally using a material testing machine (Mini Bionix II, MTS, USA, Figure 1). The peak forces represented different daily activities [Bergmann, 2010]: walking (2.3 kN), stair climbing (4.3 kN) and stumbling (5.3 kN). 2,000 loading cycles (f=1 Hz) were applied for each load level. Six eddy-current sensors, placed between stem and head, were used to determine the displacement (interface micro-motion and elastic deformation) between head and stem (Figure 1). A finite element model (FEM) based on CAD data was used to determine the elastic deformation of the prostheses for the experimentally tested activities (Abaqus, Simulia, USA). Tie-junctions at all interfaces prevented relative movements of the adjacent surfaces. The resultant translations at the centre of the ball head were determined using a coordinate transformation and a subsequent subtraction of the elastic deformation.

Materials and Methods

Two ceramic liners were instrumented at the outer contour with five strain gauge (SG) rosettes each (Fig.1). First, metal shells were axially seated in an asymmetric press-fit model with 0.5 mm under-reaming, then liners were assembled with a 2 kN axial load. SG5 was placed at the flat area of the liner, the other four were placed circumferentially in 90 degrees offset on the rear side. SG2 and SG4 were mounted opposite to each other in press-fit direction while SG1 and SG3 were placed in the non-supported direction. Three inclination angles (0°, 30°, 45°) were tested under in vivo relevant loads of 4.5 and 11 kN. Four positive angular gaps (A1=0.162°±0.007°, A2=0.084°±0.002°, A3=0.054°±0.004°, A4=0.012°±0.005°) and one negative angular gap (A5=−0.069°±0.006°) were examined. For all tests a mid-tolerance clearance between liner and ball head of 70 µm was chosen. Strain data were converted to stresses and compared using a paired 2-sided Wilcoxon Signed Rank Test at an α-level of 0.05.

Materials and Methods

The geometry of taper contact surfaces was determined using a Coordinate Measurement Machine (BHN 805, Mitutoyo, Japan). An algorithm based on the individual unworn areas of the respective taper surfaces was applied to all retrievals. One retrieval is additionally investigated by infinite focus microscopy (G4, Alicona, Austria) in the main wear areas on the neck piece taper, and the bottom, facing each other inside the junction (surfaces of the distal end of the male and the bottom of the female taper).

Methods

Cobalt chrome heads (M-SPEC, 36mm, +1.5mm; n=5) were assembled on titanium femoral stems (Corail 12/14, both components Depuy Synthes; n=5) by quasistatic axial push-on forces (F=0.5kN, 1kN, 2kN). Heads were modified by milling a flat plane, to which the joint load was applied alternately to point A and point B for 20 cycles to provide reversing moments (heel-strike FA=1971N, MA=5.4Nm; toe-off FB=807N, MB=4.6Nm; Fig. 1). All 6 degrees of freedom of relative displacement between head and stem were determined in the unloaded state and after each loading cycle. A coordinate measurement machine (accuracy ±2µm) was used to determine the components positions. Pull-off forces were measured after the last loading cycle. Each taper was tested in pristine condition and then contaminated with a bone chip (1.7±0.2mg).

Methods

One design of a dual modular hip prosthesis (Ti6Al4V, Metha®, Aesculap AG, Germany) with a high offset neck adapter (CoCrMo, CCD-angle of 130°, neutral antetorsion) and a monobloc prosthesis (stem size 4) of the same implant type were used to characterize the metal ion release of modular and non-modular hip implants. Stems were embedded in PMMA with 10° adduction and 9° flexion according to ISO 7206-6 and assembled with ceramic (Biolox® delta) or CoCrMo femoral heads (XL-offset) by three light impacts with a hammer. All implant options were tested in four different test fluids: Ringer's solution, bovine calf serum and iron chloride solution (FeCl3-concentration: 10 g/L and 114 g/L). Cyclic axial sinusoidal compressive load (Fmax = 3800 N, peak load level of walking based on in vivo force measurements [3]) was applied for 10 million cycles using a servohydraulic testing machine (MTS MiniBionix 370). The test frequency was continuously varied between 15 Hz (9900 cycles) followed by 1 Hz (100 cycles). The metal ion concentration (cobalt, chromium and titanium) of the test fluids were analysed using ICP-OES and ICP-MS at intervals of 0, 5·105, 2·106 and 10·106 cycles (measuring sensitivity < 1 µg/L).

Materials and Methods

Two ceramic liners were instrumented at the outer contour with five strain gauge (SG) rosettes (measuring grid length: 1.5 mm) on each liner (Fig.1). Metal shells were seated in an asymmetric press-fit Sawbones® model using a 0.5 mm under-reaming, and liners were afterwards axially assembled with a 2 kN load. SG5 was placed at the flat area of the liner, the other four were placed circumferentially in 90 degrees offset on the rear side of the liner. SG2 and SG4 were mounted opposite to each other in press-fit direction (contact of metal shell to the Sawbones® block) whereas SG1 and SG3 were placed in the non-supported direction (no contact of metal shell to the Sawbones® block). Four different inclination angles (0°, 30°, 45°, 60°) were tested under in vivo relevant loads of 4.5 and 11 kN. Two ceramic ball heads were used to examine a mid tolerance clearance and a clearance at the lower tolerance limit. Strain data was converted to stresses and compared using a paired two-sided Wilcoxon Rank Sum Test at an α-level of 0.05.

The tribological aspect

Hard-on-hard materials enable mixed or fluid film lubrication due to their good wettability. The development of a fluid film layer is encouraged by smaller surface pressures (larger area) and higher velocity at the articulating interface (larger radius), suggesting that larger diameters exhibit better lubrication and such less wear. This was effectivly proven in pre-clinical simulator studies and used as argument to increase the diameters of metal-on-metall and ceramic-on-ceramic bearings. Clinically the tribological advantage of larger diameters has not yet been shown. For hard-on-soft bearings the situation is different. Due to the bad wettability of Polyethylene (PE), the abrasive wear regime is dominant. This means that the longer wear path of a larger diameter will inevitably carry a larger amount of wear debris. Despite this relation, the heads used in combination with PE were also increased up to 40mm diameter, justified by the overall greatly reduced wear amount of the new generation(s) of cross-linked PE and favourable simulator results. First in-vivo studies have shown that larger heads carry larger amounts of wear particles. Whether this increase is relevant with respect to osteolysis is still unclear and will have to be shown in longer term studies.

The biomechanical aspect

Larger heads require a larger “jumping” distance until they dislocate. Consequently the use of larger heads reduces dislocation rates, which was shown in multiple clinical studies. However, the reduction in dislocation rate achieved by increasing diameters varies greatly. Some centres achieve dislocation rates below 1% with 28mm heads, other centres require 36mm heads to achieve the same result. No study shows any further advantage with head diameters larger than 36mm. Despite their obvious biomechanical advantage with regard to stability, larger heads also have large disadvantages. Larger heads carry inevitably larger friction moments, requiring better anchoring of the components. In unfavourable conditions (start-up, break-down of lubrication film), friction moments of hard-on-hard bearings can get very high and reach or even exceed the losening torque of the head on the taper. Depending on the head impaction foce during assembly, the loosening torques amount to 8 to 17Nm. Movement at the head-taper connection possibly causes wear and increased corrosion at this interface. Larger head diameters also require thinner shells and/or liners, leading to problems with liner chipping or incomplete seating. Large head diameters have also lead to the use of sub-hemispherical cups with reduced covering surface, increasing the risk of fluid film break down due to edge loading if not well positioned. Finally, larger heads might give the surgeon a wrong feeling of security regarding a sub-optimal positioned cup.

The question regarding “the optimal” head diameter is open for discussion and needs to consider the bearing material used. Head size should be limited to a reasonable compromise, which based on the information currently available, could be 36mm. Join the “36 and under” club.