Several factors influence the healing process and the long-term mechanical stability of cementless fixed implants, such as bone remodeling and mineralization processes. Histomorphometric and bone hardness measurements were taken in implants inserted in sheep femoral cortical bone at different times to compare the in vivo osseointegration of titanium screws (diam.; 3.5 x 7 mm length) with the following surface treatments: machined (Ti-MA); acid-etched (Ti-HF); HA vacuum plasma spray (Ti-HA); and Ca-P anodization followed by a hydrothermal treatment (Ti-AM/HA). Ti-MA and Ti-AM/HA implants presented the lowest (Ra = 0.20 +/- 0.01 microm) and highest (Ra = 1.97 +/- 0.64 microm) significant (p < 0.0005) roughness value, respectively. Bone-to-implant contact of Ti-HF was lower than that of the other surface treatments at both experimental times (8 weeks: -20%, ns; 12 weeks: -30%, p < 0.01). Significant differences in MAR (mineral apposition rate) were also found between the different experimental times for Ti-MA (115%, p < 0.01) and Ti-HF (57%, p < 0.01), demonstrating that bone growth had slowed inside the screw threads of Ti-HA and Ti-AM/HA after 12 weeks. No bone microhardness changes in preexisting host bone were found, while Ti-MA showed the lowest value for the inner thread area at 8 weeks (HV(200 microm)= 49.8 +/- 3.8 HV). These findings confirm that osseointegration may be accelerated by adequate surface roughness and bioactive ceramic coating such as Ca-P anodization followed by a hydrothermal treatment, which enhance bone interlocking and mineralization.