Pre-clinical, bench-top assessment of Total Knee Replacements (TKR) can provide information about the inherent constraint provided by a TKR, which does not depend on the condition of the patient undergoing the arthroplasty. However little guidance is given by the ASTM standard on test configurations such as medial-lateral (M:L) loading distribution, flexion angle or restriction of secondary motions. Using a purpose built rig for a materials testing machine, four TKRs currently in widespread clinical use, including medial-pivot and symmetrical condyle types, were tested for anterior-posterior translational constraint. Compressive joint loads from 710 to 2000 N, and a range of medial-lateral (M:L) load distributions, from 70:30% to 30:70% M:L, were applied at different flexion angles with secondary motions unconstrained. It was found that TKA constraint was significantly less at 60 and 90° flexion than at 0°, whilst increasing the compressive joint load increased the force required to translate the tibia to limits of AP constraint at all flexion angles tested. Additionally when M:L load distribution was shifted medially, a coupled internal rotation was observed with anterior translation and external rotation with posterior translation. This paper includes some recommendations for future development of pre-clinical testing methods.