On the aetiology of aseptic loosening in joint arthroplasties, and routes to improved cemented fixation

Abstract: Background: Aseptic loosening is the most common single cause for revisions in major joint arthroplasties worldwide. Different theories for aseptic loosening have been discussed during the last two decades. There is a need for improving fixation in both cemented and uncemented implants. Aims: To investigate if wear particles alone can induce aseptic loosening in a stable joint implant and if sodium fluoride loaded bone cement improves fixation of implants in rabbits with or without oestrogen defiency.Material, Methods and Results: The Miller-Galante I (MG I) prosthesis was designed to achieve permanent stability through bone ingrowth into a titanium fiber mesh. Thirty-five knees in 30 patients with MG I knee replacements have been followed clinically and radiologically with a mean follow-up of 12 years. Twenty knees were revised. Light microscopical histomorphometrical investigations of tissue integration were performed on the revised components. Patello-femoral problems, especially avulsion of the polyethylene from the metal backed patella and in some cases severe metallosis have been the main reasons for revision. Substantial wear debris was exposed to the interfacial tissues. Despite this the fixation of the components has been excellent with a high degree of osseous ingrowth displayed at histological analysis. The used rabbit model was a weight bearing, articulating prosthetic joint repeatedly injected with submicron particles of Ultra-High-Molecular-Weight-Polyethylene (UHMWPE) produced in a hip simulator. Twelve lop-eared rabbits were included in the study. Histomorphometrical investigations demonstrated no statistically significant differences between rabbits receiving UHMWPE particles (test) and those that received saline (control). The Ultra-High-Molecular-Weight-Polyethylene debris did not induce any significant osteolysis, indicating that an osseointegrated implant with a sealed interface may not be affected by UHMWPE debris nor progress to aseptic loosening. We investigated if addition of sodium fluoride to poly-metylmetacrylate (Cemex®) could improve fixation to bone. Nine mature lop-eared rabbits were included in the study. A femoral prosthesis was inserted in both knees to resurface the femuropatellar articulation. The same type of Cemex® bone cement with and without NaF was used on both sides. Two screw shaped implants machined from cured rods of test and control cement were also inserted bilaterally into the proximal tibia. Qualitative and quantitative histomorphometrical studies of the bone tissue surrounding the cement in the femur and of the the intact implants revealed similar results regardless if NaF had been added or not. The removal torque necessary to loosen the implants from the bone bed did not significantly differ between the two cements. Our findings indicate that addition of NaF has little effect on implant stability and bone remodelling in healthy rabbits in the short-term perspective. Finally, we investigated bone integration of threaded implants made of cured polymethylmethacrylate containing sodium fluoride or commercially pure titanium (c.p.ti.) grade 1 in normal and oestrogen deficient New Zealand white rabbits. Nine rabbits had been ovariectomized through laparoscopy and nine served as controls. Four weeks after the ovariectomy two threaded implants made of cured bone cement with or without sodium fluoride addition were inserted in each tibia and one threaded c.p.ti grade 1 implant was inserted in each patello-femoral joint. Measurement of the peak removal torque necessary to loosen the implants and light microscopical histomorphometrical investigations of tissue integration were performed. In the ovariectomized rabbits addition of sodium fluoride to the cement resulted in an increased area of bone in the threads (p=0.04). In the non ovariectomized rabbits no significant differences were found between test and control implants. The removal torque was lower in the ovariectomized rabbits compared to the non-ovariectomized rabbits when comparing implants with sodium fluoride addition (p=0.02). The bone tissue response and the removal torque of the titanium implants were not influenced by ovariectomy in these rabbits. Conclusions: Aseptic loosening could not be demonstrated despite high loads of debris indicating that wear particles alone do not cause aseptic loosening and that additional factors are needed. A stable osseointegrated implant in rabbits had high resistance to endotoxin free submicron UHMWPE wear particles and such particles were not able to cause aseptic loosening alone. Sodium fluoride addition to commercially available bone cement may improve bone formation around the cement, although this was only demonstrated in ovariectomised rabbits.