The aim of this study was to determine the prognostic value of preoperative patellofemoral osteoarthritis, BMI and age for implant survival of unicompartmental knee arthroplasty (UKA) performed in patients meeting strict admission criteria. The data and radiographs of 437 unilateral Oxford phase III procedures (Biomet, Bridgend, UK) were analysed. All procedures were carried out or supervised by 13 specialised knee surgeons in three different hospitals. The study group comprised 437 patients with a median follow of 2.6 years (0.1-7.9). The cumulative standard case survival rate at 5 years, when there were still 101 patients at risk, was 84.7% (CI-95%: 80.1-89.3%). Young age (<60 years) was associated with a 2.2-fold increased adjusted risk of revision (CI: 1.08-4.43; p=0.03). The preoperative presence of radiological features of patellofemoral osteoarthritis was associated with a 0.3-fold reduced adjusted risk of revision (CI: 0.11-0.89; p=0.03). BMI>30 kg/m(2), gender, the surgeon performing the operation (either as an individual or categorised by annual surgical UKA caseload, i.e., more or less than 10 UKAs) and the hospital in which surgery took place did not predict implant survival of UKA. We conclude that young patients (<60 years) experience an increased early risk of revision for UKA when compared to older patients (>60 years). Obesity (BMI>30 kg/m(2)) and preoperative patellofemoral osteoarthritis are not associated with a decreased implant survival and therefore should not be considered risk factors in this context.