Purpose <p>Computer-assisted systems improve conventional instrumentation in normalizing the axis of the limb in frontal deformities. The performance of a postoperative X-ray has been questioned to verify this axis, assuming that it can be replaced by the final image that appears on the system monitor. The objective of this work is to verify whether both axes are similar.</p> Methods <p>In 123 patients with varus–valgus frontal deformity, a total knee arthroplasty was implanted with the aid of the same navigation system without previous images. On the one hand, the final femorotibial angulation that appeared on the monitor screen, and on the other hand, the one observed in a long postoperative digitalized radiograph were recorded.</p> Results <p>A discrepancy and poor correlation were observed between the femorotibial alignment verified radiographically and that shown on the system monitor at the end of the surgical procedure. Differences were significant in the valgus deformity group (<i>p</i> = 0.04) and especially in the varus cases (<i>p</i> &lt; 0.001). The median postoperative axis as measured by computer-assisted was very close to neutral, but a deviation toward varus was observed on the X-ray.</p> Conclusion <p>The axis of the limb observed with a computer-assisted system is closer to neutral alignment than that shown on the digitalized X-ray image after total arthroplasty in patients with frontal deformities. A postoperative X-ray is recommended as certification of the result, being more reliable than the final image from the system monitor.</p>