A systematic review of the English language literature has suggested that the performance of linked and unlinked elbow replacement implants differ in terms of function, survival and mode of failure; however, in this review, only one comparative series using contemporary implants was identified. We have performed a cohort study of Kudo, Souter-Strathclyde and Coonrad-Morrey elbow replacements performed at a single centre by or under the direct supervision of a single Consultant shoulder and elbow surgeon to see if these findings were reflected in clinical practice. The first forty implantations in patients with Rheumatoid arthritis for each device have been reviewed with respect to surgical complications, elbow function and implant survival. The follow-up was shorter for the Coonrad-Morrey cohort. In terms of pain relief and range of motion, the performance of the implants was comparable. The mode of failure was different, with no dislocations/ instability seen with the linked Coonrad-Morrey implants. The loosening rate of the Coonrad-Morrey implants (both clinical and radiographic) was lower, albeit with a shorter follow-up period. The loosening rates seen in this series were higher than those previously reported in the English language literature. We conclude that the functional performance of the implants, at similar stages of the surgical learning curves, are similar in patients with Rheumatoid arthritis, but that use of a linked implant removes the risk of post-operative instability and may reduce the risk of the radiographic and clinical loosening.