This paper reports analysis obtained from 200 implant cases retrieved from humans and submitted to the American Academy of Implant Dentistry Research Foundation, Medical College of Georgia implant retrieval center. The samples that were not decalcified were embedded in polymethylmethacrylate and examined with scanning electron microscopy and routine light, polarized, or Nomarski microscopy. Cases included both orthopedic and dental implants, as well as entire mandibles and portions of maxillae obtained at autopsy. A significant number of submitted implants had substantial amounts of adhered bone, which permitted evaluation of human bone remodeling to osseointegrated implants. These implants failed because of implant fracture. As was observed with animal studies, healthy bone supported these implants, with the bone containing an interdigitating canaliculi network that provided communication between interfacial osteocytes and osteocytes deeper within the remodeled osteonal and trabecular bone. Early dental implants containing a coating of beads showed a connective tissue interface, which corresponded to the bead surface of specific orthopedic implants that underwent some degree of micromovement. This is in contrast with the excellent response reported for successful contemporary beaded implants. Significant numbers of osseointegrated fractured hydroxyapatite (HA)-coated dental implants demonstrated the adequate serviceability of these implants before biomaterial fracture. In contrast, the HA coating was dissociated from retrieved orthopedic implants, leading to extensive cup loosening and case failure. This study, therefore, underscores the need for evaluation of failed human dental and orthopedic implants. Correlations can be drawn between human retrieval and experimental animal studies.