PurposeThe purpose of this study was to understand why the revision rate of unicompartmental knee replacement (UKR) in the National Joint Registry (NJR) is so high. Using radiographs, the appropriateness of patient selection for primary surgery, surgical technique, and indications for revision were determined. In addition, the alignment of the radiographs was assessed.MethodsOxford UKR registered with the NJR between 2006 and 2010 and subsequently revised were identified by the NJR. A blinded review was undertaken of pre-primary, post-primary, and pre-revision anteroposterior and lateral radiographs of a sample of 107 cases from multiple centres.ResultsThe recommended indications were satisfied in 70%, with 29% not demonstrating bone-on-bone arthritis. Major technical errors, likely leading to revision, were seen in 6%. Pre-revision radiographs were malaligned and, therefore, difficult to interpret in 53%. No reason for revision was seen in 67%. Reasons for revision included lateral compartment arthritis (10%), tibial loosening (7%), bearing dislocation (7%), infection (6%), femoral loosening (3%), and peri-prosthetic fracture (2%, one femoral, one tibial).ConclusionsOnly 20% of the revised UKR were implanted for the recommended indications, using appropriate surgical technique and had a mechanical problem necessitating revision. One-third of primary surgeries were undertaken in patients with early arthritis, which is contraindicated. Two-thirds were presumably revised for unexplained pain, which is not advised as it tends not to help the pain. This study suggests that variable and inappropriate indications for primary and revision surgery are responsible for the high rates of revision seen in registries.Level of evidenceIII, Therapeutic study.