This study compared osteoblast proliferation, differentiation, and protein synthesis on new and used titanium (Ti) disks to test the hypothesis that cleaning and resterilization of previously used Ti disks does not alter cell response to a particular surface. Ti disks of varying roughness were prepared by one of five different treatment regimens. Standard tissue culture plastic was used as a control. Human osteoblast-like cells (MG63) were cultured on the Ti disks and cell proliferation, cell differentiation, RNA synthesis and matrix production (collagen and noncollagen protein; proteoglycans) measured. After their first use, the disks were cleaned, re-sterilized by autoclaving. and MG63 cells cultured on them as before. At confluence, the same parameters were measured and cell behaviour on new and used disks compared. When confluent cultures of cells on plastic were compared to those cultured on new Ti surfaces, cell number was reduced on the roughest surfaces and equivalent to plastic on the other surfaces. Cell number was further reduced when disks with the roughest surfaces were re-used; no differences in cell number could be discerned after cleaning and re-sterilization. Cell proliferation was inversely related to surface roughness and was less than seen on tissue culture plastic. Re-use of the Ti disks resulted in no change in cell proliferation rate. Alkaline phosphatase specific activity in isolated cells was lowest on the rougher surfaces; no differences between new and used disks were observed. Similarly, enzyme activity in the cell layer was decreased in cultures grown on rougher surfaces, with no effect of prior disk use being noted. RNA synthesis was decreased with respect to plastic in cultures on smoother surfaces and increased on rougher surfaces; prior disk use did not alter RNA synthesis. Collagen production by the cells was decreased on smoother surfaces, but was comparable to tissue culture plastic when grown on rougher surfaces. Non-collagen protein production was unaffected by culture surface and whether or not the disk had been previously used. Proteoglycan synthesis by cells was decreased on all surfaces studied and comparable on both new and used disks. The results of this study indicate that Ti implant surfaces are unaffected by cleaning and resterilization, although rougher surfaces may require more extensive cleaning than smoother ones. This suggests the possibility that implants, in the same patient, could be safely reused. In vivo studies in animals, however, need to be performed before clinical application can be considered.