Femoral component malalignment is one of the main causes of persisting anterior knee pain after knee replacement. This study examined interindividual reproducibility in perioperative definition of the transepicondylar axis (TEA) as a reference for measuring the rotational alignment of the femoral component. Eight surgeons experienced in knee prosthetic surgery marked on Thiel-embalmed cadaver specimens the reference points that they would normally use to define the TEA during knee replacement. These were digitized by a video system, and all the spots defined by the surgeon were translated into a reference picture, allowing a digital analysis of the distances between all the spots marked. The maximal distance between the spots that the participants had marked as relevant for the TEA was 13.8 mm at the lateral and 22.3 mm at the medial epicondyle. Projecting all spots marked into one picture resulted in an area of 116 mm2 on the lateral and 102 mm2 on the medial epicondyle. The median range of the fault between two different participants was 6.4 mm on the lateral side (range 13.2 mm) and 9.7 mm on the medial (range 21.6 mm). Because the rotational alignment of the femoral component is extremely relevant for successful implantation of total knee prosthesis, the interindividual discrepancy in defining the TEA as reference is rather high. As this reference line is commonly used, the perioperative variance and the resulting rotational discrepancy of the femoral component must be considered.