BACKGROUND. Despite extensive current knowledge about fractures of the femoral shaft, the choice between antegrade and retrograde intramedullary (IM) nailing with respect to the future function of the joint serving to introduce the nail continues to raise controversy. To compare knee function in patients with a healed fracture of the femoral shaft fixed by antegrade vs. retrograde IM nailing. MATERIAL AND METHODS. The study involved a group of 65 individuals with traumatic fractures of the femoral shaft who underwent stabilisation with IM nails in the years 2001-2010. Thirty-two cases were retrograde nails (Group R) and 33 antegrade nails (Group A). Patient age at trauma ranged from 19 to 91 years (mean: 47). Knee function was assessed in both groups with the KOOS, KSS1 and KSS2 scoring systems. RESULTS. Knee function as assessed with KOOS differed significantly between retrograde and antegrade nailing, with a greater incidence of poor and fair results in the former and more excellent outcomes in the latter group (p=0.0133). As regards KSS1 and KSS2, there were no significant differences between the groups (p=0.1947, p=0.4038). The range of motion was 86-125 degrees in Group R and 121-125 degrees in Group A. Knee pain was reported by 37.5% of the patients treated with retrograde nailing and 39.4% of those who had the IM nail inserted via the antegrade approach (p=0.22). The mean time to bone union was 180 days in Group R and 219 days in Group A (p=0.25). Age and presence of osteoarthritis at trauma significantly lowered the KOOS (p=0.0027, p= 0.005) and KSS (p=0.0002, p=0.002) scores, as well as the knee range of motion (p=0.0014, p=0.004) CONCLUSIONS. 1. Knee function following retrograde and antegrade IM nailing to stabilise femoral shaft fractures was comparable. 2. The choice of IM nailing method should not be based solely on orthopaedic indications, but also on the severity of osteoarthritis present at trauma.