How correctly does an intramedullary rod represent the longitudinal tibial axes?

In a robot-assisted procedure for preparing the tibia in total knee arthroplasty, developed in the authors' laboratory, an intramedullary rod is used to register the tibia. In 18 formalin-fixed tibias, the difference in orientation was calculated between the intramedullary rod and several longitudinal tibial axes used in clinical practice. This was done using roentgenstereophotogrammetric analysis. Three tibial axes and two insertion techniques were considered. In three-dimensional space, small differences between the axes are observed. The results showed a high standard deviation, indicating the importance of anatomic differences. In the frontal plane, the difference in orientation between rod and tibial axes never exceeded +/- 2 degrees. In the sagittal plane, the observed differences were larger. Significant differences between the considered axes appeared. The results of the two insertion techniques were not significantly different. Because an intramedullary rod frequently is used for alignment of the tibia in conventional surgery, these results also are valuable for conventional surgery. In the current study, the accuracy of the intramedullary alignment is examined, without influences of the sawing procedure. Moreover, the study is not limited to the frontal plane; the total accuracy in three-dimensional space, and the accuracy in the frontal and the sagittal planes were studied.