Abstract
Introduction
From pre-operative planning to final implant cementation, total knee arthroplasty (TKA) preparation is a succession of many individual steps, each presenting potential sources of error that can result in devices being implanted outside the targeted range of alignment. This study assessed alignment discrepancy occurring during different TKA steps using an image-free computer-assisted orthopaedic surgery (CAOS) guidance system (Exactech GPS, Blue-Ortho, Grenoble, FR) in normal and abnormal mechanical axis.
Materials and methods
We used a commercially available artificial leg (MITA trainer leg M-00058, Medical Models, Bristol, UK) able to receive (neutral / varus / valgus) knee inserts simulating the proximal tibia and distal femur. A pre-surgical profile was established to define resection parameters for the proximal tibial and distal femoral cuts (Figure 1A). Data from the guidance system were collected at three separate steps: (1) cutting block adjusted but not pinned to the bone (Figure 1B), (2) cutting block adjusted and pinned to the bone (Figure 1C), and (3) after the cuts were checked (Figure 1D). These data were then compared to the resection target parameters to track potential dispersions occurring during the process.
Due to the amount of data (i.e., four studied resection parameters per bone, three operative steps, and three knee model types), the authors introduced an “error index”, which was a unitless indication of overall error magnitude obtained by averaging the absolute values of all linear and angular measurement errors. Due to knee model dimensions (∼55 mm), the authors equally considered linear and angular measurement values (i.e., 1 mm equivalent to 1°).
Results
Regardless of resection parameter or bone deformity type, all linear or angular error distributions were symmetrical around the neutral value, which implies no obvious skew in terms of error direction. The type of knee model deformity had almost no effect on overall error magnitudes throughout all surgical steps (Figure 2).
Discussion
Few studies present possible causes for errors when using CAOS for TKA. Notably, Bathis et al. evaluated cutting errors as the difference between the primary cutting block position and the resulting resection plane. As a result, errors due to a malpositioning of the guide jig itself were not described.1
In general, the authors found the dispersions at each step to seemingly be random. For both the tibia and the femur, a significant increase in the error index from the adjusted to the attached step (p<0.001 and p=0.005; respectively) was observed, meaning the pinning of the cutting block to the bone is a key step. Also, observing the relationship between linear and angular parameters was relevant. For example, for the femur, a cut in extension was highly correlated with lower than expected distal femoral resection (Pearson correlation factor of 0.783 and 0.913 at the checked step for the medial and lateral distal femoral resections; respectively, p<0.001).
Regardless of the presence and type of deformity, the evaluated image-free computer-assisted guidance system did not exhibit substantial alignment dispersions during any step of the procedure.
