Polyether-ether-ketone (PEEK) is one of the most common materials used for load-bearing orthopaedic devices due to its radiolucency and favorable mechanical properties. However, current smooth-surfaced PEEK implants can lead to fibrous encapsulation and poor osseointegration. This study compared the in vitro and in vivo bone response to two smooth PEEK alternatives: porous PEEK and plasma-sprayed titanium coatings on PEEK. MC3T3 cells were grown on smooth PEEK, porous PEEK, and Ti-coated PEEK for 14 days and assayed for calcium content, osteocalcin, VEGF and ALP activity. Osseointegration was investigated by implanting cylindrical implants into the proximal tibiae of male Sprague Dawley rats for 8 weeks. Bone-implant interfaces were evaluated using μCT, histology and pullout testing. Cells on porous PEEK surfaces produced more calcium, osteocalcin, and VEGF than smooth PEEK and Ti-coated PEEK groups. Bone ingrowth into porous PEEK surfaces was comparable to previously reported porous materials and correlated well between μCT and histology analysis. Porous PEEK implants exhibited greater pullout force, stiffness and energy-to-failure compared to smooth PEEK and Ti-coated PEEK, despite Ti-coated PEEK exhibiting a high degree of bone-implant contact. These results are attributed to increased mechanical interlocking of bone with the porous PEEK implant surface. Overall, porous PEEK was associated with improved osteogenic differentiation in vitro and greater implant fixation in vivo compared to smooth PEEK and Ti-coated PEEK. These results suggest that not all PEEK implants inherently generate a fibrous response and that topography has a central role in determining implant osseointegration.