Abstract
Introduction
Oxide-based alumina (Al2O3) is used to manufacture femoral heads for total hip arthroplasty (THA). Silicon nitride (Si3N4) is a non-oxide ceramic used to make spinal implants. Ceramic materials are believed to be bioinert, (i.e., stable under hydrothermal conditions). Indeed, clinical data have shown 15–20 year longevity of Al2O3 bearings in THA. In this work, we examined the surfaces of Al2O3 and Si3N4 after exposure to physiologic conditions to see if these ceramics are truly inert.
Materials and Methods
Four self-mated Ø28 mm diameter Al2O3 femoral heads (n=2 each of BIOLOX®forte, CeramTec, Plochingen, Germany; and BIOCERAM®, Kyocera Co., Kyoto, Japan), were retrieved during revision THA, between 7.7–10.7 years post-implantation. To simulate in vivo material aging, comparable, new Al2O3 and Si3N4 femoral heads (AMEDICA Corporation, Salt Lake City, UT, USA) were exposed to autoclave conditions (100°C-121°C; 300 hrs; n=3 heads, per material). Advanced Raman and cathodoluminescence spectroscopy, and electron microscopy were used to examine surface characteristics of each specimen, and quantify oxygen ion vacancy formation and composition.
Results
The naturally hydroxylated free surfaces of retrieved Al2O3 demonstrated lattice dissociation. Dehydroxylation of Al2O3 produced hydroxyls and proton radicals, which in turn promoted the formation of surface vacancies for preserving electrical charge neutrality. During frictional sliding of Al2O3, the continuous formation, subsequent adsorption, and frictional removal of hydroxylated sites repeatedly created and annihilated a population of different oxygen-vacancy sites. Moreover, protein by-products and in vivo released ions sub-valent with respect to Al3+ (e.g., Ca2+, Mg2+, and Na+) were found to preferentially link to oxygen-vacancy sites, inducing irreversible stoichiometric alterations of the Al2O3 surface. Oxygen vacancy formation was seen in all samples, (i.e., all retrievals and all in vitro hydrothermally exposed samples). Cation substitution and spinel formation were only observed in retrievals because these cations are not available during in vitro testing, (i.e., Ca2+, Mg2+ and Na+ come from the synovial fluid). In contrast, Si3N4 surfaces showed no evidence of direct hydrogen bond formation, and therefore no dissociative interaction with water molecules, when subjected to accelerated aging conditions (Fig. 1).
Discussion
Because of its molecular polarity, water can dissolve ionic substances, since the ionic compound interacts with either side of the water dipole. This phenomenon leads to dissociation of the ionic molecule by dehydroxylation. Our results show that Si3N4 surfaces are stable in hydrothermal environments. In contrast, Al2O3 surfaces demonstrate surface changes under in vivo and in vitro conditions because of modifications of the lattice structure (e.g., vacancy generation and formation of a soft MgAl2O4 spinel phase), which alters local mechanical properties and tribological wear resistance. Cations, such as Na+ are released from dilution of sodium hyaluronate; a phenomenon that occurs in patients with rheumatoid arthritis.
Conclusion
These data suggest that surface dehydroxylation may lead to the long-term in vivo degradation of Al2O3 bearings in THA. Covalently-bonded bioceramics, such as Si3N4 are impervious to such degradation. Si3N4 may be truly bioinert in vivo, ensuring multi-decade durability and superior performance of orthopaedic bearings.
