Purpose: To compare the clinical outcome of single, partial and complete fixed implant supported prostheses immediately loaded (within 48 h), early loaded at 6 weeks, and conventionally loaded at 3 months (delayed loading). Materials and methods: A total of 54 patients (18 requiring single implants, 18 partial fixed prostheses, and 18 total fixed cross-arch prostheses) were randomised in equal numbers at two private practices to immediate loading (18 patients), early loading (18 patients), and conventional loading (18 patients) according to a parallel group design with three arms. To be immediately or early loaded, implants had to be inserted with a torque superior to 40 Ncm. Implants were initially loaded with provisional prostheses, replaced after 4 months by definitive ones. Outcome measures were prosthesis and implant failures, complications and peri-implant marginal bone levels. Results: Two conventionally loaded patients rehabilitated with cross-arch fixed total prostheses dropped-out up to 1 year post-loading. No implant or prosthesis failed and three complications occurred, one in each loading group. Peri-implant marginal bone loss was 0.19 ± 0.44 mm at immediately loaded implants, 0.18 ± 0.66 mm at early loaded implants and 0.25 ± 0.28 mm at conventional loaded implants. There were no statistically significant differences in complications (P = 1.000) and bone loss (P = 0.806) between the three loading strategies. Conclusions: All loading strategies were highly successful and no differences could be observed for implant survival and complications when loading implants immediately, early or conventionally.