In a randomized prospective study, two implant systems were compared in forty consecutive patients treated for mandibular edentulism. The patients were randomly allotted for treatment by the Brånemark two-stage (submerged) system (BRS), or the ITI(R) one-stage (non-submerged) system. In all, 102 Brånemark selftapping implants and 106 ITI hollow screw implants were installed and all patients were treated with full bridges. Biological and prosthodontic parameters, complications, success rates, clinical efficacy, patient satisfaction and resource requirements were evaluated. No differences were found in plaque accumulation, bleeding or complications during the follow-up period. The BRS group showed deeper periimplant sulcus, less attached mucosa, larger bridge-mucosa distance and higher Periotest values. Prosthetic complications were not related to the configuration of the implant systems. After 3 years, the cumulative success rates were 97.9% and 96.8% for the Brånemark and ITI systems, respectively (difference not statistically significant). One implant in the BRS group had failed to osseointegrate at the time of abutment connection, and another was lost after 2 years due to progressive breakdown of bone. In the ITI group, three implants showed progressive bone loss after 1-3 years associated with periimplant infection. All 40 bridges were intact and remained stable throughout the study. There was general patient satisfaction, but about half the Brånemark patients reported difficulty in coping with the surgical procedures. Treatment time was similar for the two systems. It is concluded that both systems meet the current requirements for dental implant systems in the treatment of mandibular edentulism.