Background: Total elbow arthroplasties (TEAs) are traditionally grouped into linked and unlinked design. The aim was to analyze the difference in clinical outcomes after TEA based on implant design and indication for surgery and to evaluate primary and revision TEAs. Methods: A total of 167 TEAs (126 primary and 41 revision TEAs) in 141 patients were evaluated with patient-reported outcome measure by the Oxford Elbow Score (OES) and clinically assessed with the Mayo Elbow Performance Score (MEPS), range of motion (ROM), and standard radiographs. Results: The mean follow-up was 10.5 years for primary and 7.5 years for revision TEAs. There was no difference in OES or MEPS between linked and unlinked primary TEAs. The OES score in the social-psychological domain was significantly lower in TEAs performed due to fracture (67) compared with rheumatoid arthritis (81; P = .025). ROM in extension-flexion was 116° for primary linked TEAs compared with 110° for primary unlinked TEAs (P = .02). Revision TEAs were associated with a poorer outcome in OES, MEPS, and ROM compared with primary TEAs. Radiographic signs of loosening were seen in 15 primary and 7 revision TEAs at follow-up. Conclusions: We found no clinically significant differences in outcomes after linked or unlinked TEAs. Patients with TEAs due to fracture had poorer social-psychological results than rheumatoid arthritis patients. The results after revision surgery were significantly inferior compared with primary procedures. The OES contributes to the evaluations of the outcome after TEA surgery with a nuanced picture of the patient's perception.