Differential response of Staphylococci and osteoblasts to varying titanium surface roughness

The surface roughness of metallic orthopaedic implants has typically been used to influence osseointegration and spatially control load transfer to the surrounding bone. Because of the increasing recognition of biomaterials-associated infection as a leading implant failure mode, we are interested to know the relative importance of roughness not only on surface–osteoblast interactions but also on surface–bacteria interactions. This in vitro study thus compares the effects of surface topography on Staphylococcus epidermidis and human osteoblast behavior using four clinically relevant titanium surface finishes: polished, satin, grit-blasted and plasma-sprayed. Important differences between these surfaces are manifested not only by their vertical roughness parameters but also by the lateral length scales over which topographic fluctuations occur. We find that S. epidermidis adhesion and growth is substantially higher on the satin and grit-blasted surfaces than on the polished or plasma-sprayed surfaces. The former are both substantially rougher at length scales comparable to that of bacteria. In contrast, based on imaging and biochemical assays of proliferation, differentiation and matrix formation, we find that desirable osteoblast–surface interactions are maximized on plasma-sprayed surfaces and minimized on satin-finished surfaces. We attribute these differences to the fact that the plasma-sprayed surface is relatively smooth compared to the size of an individual osteoblast, while the satin surface is rough at this length scale. These findings indicate that both the vertical and lateral character of surface roughness can be modified to not only optimize implant–bone interactions but to simultaneously minimize implant–bacteria interactions.