BackgroundOne of the remaining issues in total knee arthroplasty (TKA) is achieving sufficient rotational arc during deep flexion range of the knee for specific postures such as the ‘seiza’ or cross-legged sitting. This study aimed to evaluate whether there was a change in the actual in vivo rotational arc during deep flexion range before and after a design change of polyethylene (PE) inserts. MethodsIn 50 posterior-stabilised TKA cases, knee kinematics, including rotational movement, were measured intraoperatively using an image-free navigation system to compare a newly designed PE insert with reduced the posterior lip with a conventional PE insert. Femoral–tibial rotational angles at 30°, 45°, 60°, 90°, 120°, and 130° knee flexion were evaluated. Varus/valgus instability, knee range of motion, and femoral rollback were also measured. Obtained parameters were compared between new and conventional PE inserts. The independent factors associated with rotational arc during deep flexion range (120° and 130° knee flexion) were analysed using multivariate analysis. ResultsThe newly designed PE insert demonstrated a significant increase in the rotational arc at 120° (22.9 ± 8.7° vs. 30.1 ± 11.9°, P < 0.001) and 130° (24.3 ± 9.5° vs. 32.5 ± 12.4°, P < 0.001) knee flexion compared with that with the conventionally designed posterior-stabilised insert. Multivariate analysis demonstrated that using the newly designed PE insert was an independent predictor of improved rotational arc during deep flexion range: regression coefficient was 11.2 (95% confidence interval 7.1–15.3, P < 0.001). ConclusionThe design change, which reduced the posterior lip of the PE insert, contributed to improved rotational arc in 120° and 130° deep flexion ranges.