This report is based on a total of 2955 implants of 6 different designs, randomized and placed in 829 patients and followed for 48 months. Implant failure was defined as nonintegration at uncovering or removal due to mobility, persistent pain, infection, and evidence of radiographic bone loss. Failures were reported for 3 phases of treatment: implant placement to uncovering (phase 1), uncovering to loading (phase 2), and postloading (phase 3). Differences in survival were compared with Kaplan-Meier survival curves. The maxillary single tooth application resulted in 95.2% survival for the hydroxyapatite-coated grooved implants. In the maxillary completely edentulous application, survival of hydroxyapatite grooved and screw implants were considerably better compared with the titanium screw implants. The hydroxyapatite-coated cylinder had better survival than the titanium basket and screw designs in the mandibular completely edentulous application. The hydroxyapatite-coated cylinder and grooved implants in the maxillary posterior partially edentulous application had similar survival rates. The survival of the hydroxyapatite-coated cylinder exceeded that of the titanium basket in mandibular posterior partially edentulous applications. Analyses by phase of treatment indicated a pattern of early failure for nonhydroxyapatite-coated implants compared with hydroxyapatite-coated implants. The implant with the highest survival at all phases of treatment was the hydroxyapatite-coated press-fit cylinder. Two hydroxyapatite-coated implant designs performed well in the challenging posterior maxillary region.