To analyze the mechanism of failure and basic cause in cases of early failure which were required revision within 5 years index TKA. Between 1991 and 2006, 167 revisions TKA of aseptic failure were performed. Revision diagnosis or reason for failure were categorized as wear of tibial polyethylene insert, failure of tibial base plate, early imbalance between medial and lateral soft tension, tight or loose PCL and posterior capsule. The percentages of each failure category were calculated as a percentage of the overall number of revision TKA and a percentage of the early failures. A descriptive statistics were calculated for the time in situ for each failure category.

Early failure within 5 years following index TKA occurred in 33 out of 167 TKA(20.0%). Average time in situ was 38.53 months(3.21 years). Wear of the tibial polyethylene insert occurred in 12 out of 33 cases(36.4%). All cases showed tight PCL. Loosening was the second leading cause occurring in 9 cases(27.2%). Pure instability with tight MCL occurred in 3 knees. Catastrophic early wear within one year after index surgery occurred in 18 knees. The cause of failure were flat surfaced poly in 11, fracture of metal tray 2, dislocation of the thick poly insert 1 and early poly wear due to unknown cause 4.

There were multiple factors of the early failure, which could be divided into design failure and surgical skill failure. However, they worked together in most of the cases.

Beta–tricalciumphosphate(β-TCP)coatinglayerisknown to be resorbed much faster than hydroxyapatite(HA), however, there has been few reports explaining the exact mechanism until now. Therefore, we investigated whether the resorption mechanisms of these two compounds are same, if not, what is the difference.

Eighty titanium discs with 12mm in diameter and 2mm in thickness were coated with HA(n=40) or β-TCP(n=40) by dip and spin coating method. In each group, the specimens were divided into 2 subgroups respectively; Dissolution (D, n=20) group and Osteoclast culture (C, n=20) group. The coated discs in D group were immersed in the cell culture media for 5 days, whereas, in C group, osteoclast-like cells (5×103 cells/500μ), which were isolated form human giant cell tumor, were seeded on the specimens and cultured for 5 days. Cultured cells were defined as osteoclast by the determination of osteoclast marker (tartrate-resistant acid phosphatase, TRAP). After immersion or osteoclast culture, the dissolution characteristics of coating surface were observed using light microscope (LM) and scanning electron microscope (SEM). And the area fraction of resorption lacunae formed by osteoclast was analyzed by image analysis to evaluate the activity of osteoclastic degradation.

After 5 days of dissolution, there were much more cracks and denuded areas in β-TCP coating compared to HA coating. In C group, the osteoclasts covering the coating layer were identified on LM and SEM images. Mean area fraction of resorption lacunae in HA-C group was 11.62%, which was significantly higher than that of 0.73% of β-TCP-C group (p=0.001).

We conclude that the resorption mechanism of HA and β-TCP coating layers was different each other in vitro study. The coated β-TCP was degraded mainly by dissolution and also tended to be separated from implant, on the other hand, the HA coating layer was resorbed by osteoclastic activity