OBJECTIVE: Acute knee injury is associated with post-traumatic OA (PTOA). Very little is known about the genome-wide associations of PTOA when compared with idiopathic OA (iOA). Our objective was to describe the development of knee OA after a knee injury and its genetic associations in UK Biobank (UKB). DESIGN: Clinically significant structural knee injuries in those ≤50 years were identified from electronic health records and self-reported data in 502,409 UKB participants. Time-to-first knee osteoarthritis (OA) code was compared in injured cases and age-/sex-matched non-injured controls using Cox Proportional Hazards models. A time-to-OA genome-wide association study (GWAS) sought evidence for PTOA risk variants 6 months to 20 years following injury. Evidence for associations of two iOA polygenic risk scores (PRS) was sought. RESULTS: Of 4233 knee injury cases, 1896 (44.8%) were female (mean age at injury 34.1 years [SD 10.4]). Over a median of 30.2 (IQR 19.5-45.4) years, 1096 (25.9%) of injured cases developed knee OA. The overall hazards ratio (HR) for knee OA after injury was 1.81 (1.70,1.93), P = 8.9 × 10-74. Female sex and increasing age at injury were associated with knee OA following injury (HR 1.15 [1.02,1.30];1.07 [1,07,1.07] respectively). OA risk was highest in the first 5 years after injury (HR 3.26 [2.67,3.98]), persisting for 40 years. In 3074 knee injury cases included in the time-to-OA GWAS, no variants reached genome-wide significance. iOA PRS was not associated with time-to-OA (HR 0.43 [0.02,8.41]). CONCLUSIONS: Increasing age at injury and female sex appear to be associated with future development of PTOA in UKB, the risk of which was greatest in the 5 years after injury. Further international efforts towards a better-powered meta-analysis will definitively elucidate genetic similarities and differences of PTOA and iOA.