Wear of ultra-high molecular weight polyethylene and polytetrafluoroethylene in a hip simulator: a dose-response study of protein concentration.

Charnley's laboratory wear studies of non-gamma sterilized polytetrafluoroethylene (PTFE) and polyethylene (PE) found that the PTFE to PE wear-rate ratio of 250:1 was much higher than the in-vivo wear ratio of 20:1. Tests of PTFE and PE in our laboratory showed a wear ratio of 150:1, using bovine serum as the lubricant and 190:1 with water as the lubricant. Our hypothesis was that the wear-rates of PTFE and PE cup materials were related to the concentration of protein in the serum. We studied the wear behavior of PTFE and PE cups in varied protein concentrations, using 4 femoral head sizes to validate the clinical range reported by Charnley. The PTFE wear-rates increased with increasing protein concentration and conversely, PE wear-rates decreased with increasing protein concentration. This inverse relationship made it possible to bring the wear ratio closer to the desired clinical wear ratio. We found that the clinically relevant PTFE/PE wear ratio corresponded to 3-10 mg/mL of protein in bovine serum.

Ultra-low wear rates for rigid-on-rigid bearings in total hip replacements.

With the increased clinical interest in metal-on-metal and ceramic-on-ceramic total-hip replacements (THRs), the objective of this hip simulator study was to identify the relative wear ranking of three bearing systems, namely CoCr-polyethylene (M-PE), CoCr-CoCr (M-M) and ceramic-on-ceramic (C-C). Volumetric wear rates were used as the method of comparison. The seven THR groupings included one M-PE study, two M-M studies and four C-C studies. Special emphasis was given to defining the 'run-in' phase of accelerated wear that rigid-on-rigid bearings generally exhibit. The hypothesis was that characterization of the run-in and steady state wear phases would clarify not only the tribological performance in vitro but also help correlate these in vitro wear rates with the 'average' wear rates measured on retrieved implants. The implant systems were studied on multichannel hip simulators using the Paul gait cycle and bovine serum as the lubricant. With 28 mm CoCr heads, the PE (2.5 Mrad/N2) wear rates averaged 13 mm3/10(6) cycles duration. This was considered a low value compared with the clinical model of 74 mm3/year (for 28 mm heads). Our later studies established that this low laboratory value was a consequence of the serum parameters then in use. The mating CoCr heads (with PE cups) wore at the steady state rate of 0.028 mm3/10(6) cycles. The concurrently run Metasul M-M THRs wore at the steady state rate of 0.119 mm3/10(6) cycles with high-protein serum. In the second Metasul M-M study with low-protein serum, the THR run-in rate was 2.681 mm3/10(6) cycles and steady state was 0.977 mm3/10(6) cycles. At 10 years, these data would predict a 70-fold reduction in M-M wear debris compared with the clinical PE wear model. All M-M implants exhibited biphasic wear trends, with the transition point at 0.5 x 10(6) cycles between run-in and steady state phases, the latter averaging a 3-fold decrease in wear rate. White surface coatings on implants (coming from the serum solution) were a confounding factor but did not obscure the two orders of magnitude wear performance improvement for CoCr over PE cups. The liners in the alumina head-alumina cup combination wore at the steady state rate of 0.004 mm3/10(6) cycles over 14 x 10(6) cycles duration (high-protein serum). The zirconia head-alumina cup THR combination wore at 0.174 and 0.014 mm3/10(6) cycles for run-in and steady state rates respectively (low-protein serum). The zirconia head and cup THR combination wore slightly higher initially with 0.342 and 0.013 mm3/10(6) cycles for run-in and steady state rates respectively. Other wear studies have generally predicted catastrophic wear for such zirconia-ceramic combinations. It was noted that the zirconia wear trends were frequently masked by the effects of tenacious white surface coatings. It was possible that these coatings protected the zirconia surfaces somewhat in this simulator study. The experimental ceramic Crystaloy THR had the highest ceramic run-in wear at 0.681 mm3/10(6) cycles and typical 0.016 mm3/10(6) cycles for steady state. Since these implants represented the first Crystaloy THR sets made, it was likely that the surface conditions of this high-strength ceramic could be improved in the future. Overall, the ceramic THRs demonstrated three orders of magnitude wear performance improvement over PE cups. With zirconia implants, while the cup wear was sometimes measurable, head wear was seldom discernible. Therefore, we have to be cautious in interpreting such zirconia wear data. Identifying the run-in and steady state wear rates was a valuable step in processing the ceramic wear data and assessing its reliability. Thus, the M-M and C-C THRs have demonstrated two to three orders of reduction in volumetric wear in the laboratory compared with the PE wear standard, which helps to explain the excellent wear performance and minimal osteolysis seen with such implants at retrieval operations.

Synovial fluid from loose hip arthroplasties inhibits human osteoblasts.

Aseptic loosening of prosthetic components in patients who have undergone total hip arthroplasty is a major clinical problem. Earlier studies on this topic have focused mainly on different aspects of bone resorption. The current study investigated the influence of synovial fluid from patients who underwent revision surgery because of aseptic loosening and synovial fluid from patients with osteoarthritis on the proliferation of primary cultures of human osteoblasts. Incubation of cells with 10% synovial fluid from patients who had revision surgery significantly inhibited [3H]thymidine incorporation into deoxyribonucleic acid in human osteoblasts compared with control conditions, whereas 10% synovial fluid from patients with osteoarthritis had a significant stimulatory effect. These findings correlate well with clinical features seen in these diseases, such as increased net bone resorption around the prosthesis in patients with loosening, and increased periarticular bone formation in patients with osteoarthritis.

Charnley wear model for validation of hip simulators--ball diameter versus polytetrafluoroethylene and polyethylene wear.

Wear rates of polytetrafluoroethylene (PTFE) and polyethylene cups were compared in 9-channel and 12-channel simulators, using serum lubrication and gravimetric techniques for wear assessment. Cobalt-chromium (CoCr) and alumina ceramic femoral heads in 22-42 mm diameter range were used to validate simulator wear rates against clinical data. This was also the first study of three femoral head sizes evaluated concurrently in a simulator (with three replicate specimens) and also the first report in which any wear experiments were repeated. Fluid absorption artefacts were within +/-1 per cent of wear magnitude for PTFE and +/-8 per cent for polyethylene and were corrected for. Wear rates were linear as a function of test duration. Precision within each set of three cups was within +/-6 per cent. The wear rates from experiments repeated over 15 months were reproducible to within +/-24 per cent. However, the magnitudes of the simulator wear rates were not clinically accurate, the PTFE wear rates (2843 mm3/10(6) cycles; 22 mm diameter) were over three times higher than in vivo, the polyethylene 30 to 50 per cent on the low side (23 mm3/10(6) cycles; 22 mm diameter). Volumetric wear rate increased with respect to size of femoral head and a linearly increasing relationship of 7 8 per cent/mm was evident with respect to femoral head diameter for both PTFE and polyethylene. These data compared well with the clinical data.

Water and bovine serum lubrication compared in simulator PTFE/CoCr wear model.

Controversy surrounds wear data from laboratory hip simulator studies, whether derived from water-based or serum-based studies or whether a major design parameter such as the size of the femoral head has an effect on the volume of wear particulate released. To investigate these relationships, we studied cup wear in water- and serum-based lubricants using as our standard the polytetrafluoroethylene (PTFE) data derived by Charnley. To model Charnley's clinical experience, PTFE acetabular cups were used in sets of three each of four sizes of CoCr femoral heads: 22.25-, 28-, 32-, and 42-mm diameters. Six criteria were used to evaluate the performance of the lubricants against clinical accuracy and scientific methods. The PTFE wear data from the serum-based tests was consistently linear with the duration of the test, exhibited a precision within +/-3% about the average for each set of three cups, and copious amounts of wear debris were clearly seen circulating and settling to the bottom of the wear chambers. The wear data clearly demonstrated Charnley's thesis that volume of wear increases with regard to the size of the femoral head in a linear manner. This increase was considered satisfactory at 9%/ mm. However, in terms of clinical accuracy, the simulator wear rates averaged 3 to 4 times greater than the comparable clinical data for wear magnitude. Thus, the serum-based tests satisfied three of the six criteria used. The PTFE wear data from the water-based tests was generally nonlinear, continually increasing with test duration. These wear trends were examined in three discrete phases to estimate the changing wear rates. By the end of the tests, the wear rates had increased from 1.3 to 3.9 times, with the 42-mm heads showing the greatest change. The resulting precision was never better than +/-26% and deteriorated to +/-70%. In terms of clinical accuracy, the water-based wear rates varied from 2 to 7 times less than the Charnley PTFE wear magnitudes, averaging 4 times less. The water-based data did not satisfactorily model the relationship between increased wear with increased head size. Minimal PTFE wear debris was observed, and what did emerge after thousands of wear cycles appeared as streamers up to 30 mm long and up to 5 mm wide. When these detached, they floated up to the surface where they could be separated into smaller particulates. A similar phenomenon was noted for polyethylene wear tests conducted with water lubrication. Thus the water-based tests satisfied none of the six validation criteria evaluated. These data raise serious doubts as to the validity of testing implant and material combinations in water as a predictor of clinical performance. Bovine serum was not totally satisfactory, but the wear data did model some of the important clinical characteristics of hip joint behavior.