BACKGROUND Navigation systems have been successful in reducing the outlier of leg alignment after total knee arthroplasty (TKA). Less is known about the restoration of the anatomical joint line with computer-assisted knee replacement. The aim of this study was to determine whether joint line changes <3 or ≥3 mm are predictable with several pre- and intraoperative parameters. METHODS The study included a total of 180 cases of primary computer-assisted TKA performed using the gap-balancing/tibia-first technique. The final shift of the joint line was calculated using computer verification of proximal tibial and distal femoral cuts. In consideration of the clinical relevance of a 3-mm joint line shift, patients were stratified into two groups: Group I, with joint line change <3 mm, and Group II, with joint line change ≥3 mm. Between groups, variables such as demographics, Kellgren & Lawrence degree of osteoarthritis, preoperative flexion contracture, pre-/intraoperative mechanical leg alignment, flexion/extension gaps, and implant design/sizes were compared statistically. RESULTS The absolute joint line shift averaged 1.6 ± 1.3 mm (range 0-6 mm). A joint line shift ≥3 mm was observed in 28 cases (15 %). A statistically significant difference between groups was not identified for any of the parameters. Shift of the joint line did not correlate with the analyzed variables. CONCLUSIONS Joint line was adequately restored when computer navigation was carefully applied as a measuring tool for primary TKA. Knee joint deformity, leg alignment, gap balance, and implant type did not aid in predicting the joint line shift.