OBJECTIVE Total knee arthroplasty (TKA) is typically carried out either with retention (CR) of the posterior cruciate ligament (PCL) or with sacrifice of this ligament and implantation of a posterior stabilised (PS) prosthesis. This paper investigates a comparison of PCL function in knees treated for osteoarthritis with TKA where the PCL is preserved to those knees treated with TKA and posterior stabilisation. METHODS One hundred and sixty-eight patients (232 knees) who had undergone TKA with either a PS or CR implant were included in the study. Clinical assessment included antero-posterior (AP) laxity and posterior sag assessment with an arthrometer. RESULTS The mean AP laxity at 90° of flexion for CR TKAs was 6.5 mm (±3.1) and was the same [6.5 mm (±2.4)] as in the PS group. However, 56% of the PCL-preserved knees had a posterior sag of over 3 mm compared to 18% of the knees in the PS group. The American Knee Society Scores for either group showed that, although the mean function score was the same, the knee score was superior in the PS group (77 vs. 84). The range of motion was also superior in the PS group (111° vs. 105°). CONCLUSIONS The use of the PS prosthesis for TKA provides a more predictable outcome with regard to posterior sag and a better maximum flexion than a CR implant.