Aims. The cemented Oxford unicompartmental knee arthroplasty (OUKA) features two variants: single and twin peg OUKA. The purpose of this study was to assess the stability of both variants in a worst-case scenario of bone defects and suboptimal cementation. Methods. Single and twin pegs were implanted randomly allocated in 12 pairs of human fresh-frozen femora. We generated 5° bone defects at the posterior condyle. Relative movement was simulated using a servohydraulic pulser, and analyzed at 70°/115° knee flexion. Relative movement was