Pre-operative templating for knee arthroplasty shows low accuracy with standard X-rays

Objectives The purpose of this study was to evaluate the accuracy and reliability of pre-operative templating in predicting the size of femoral and tibial components and the effect of coronal deformity on templating accuracy. Methods This was a retrospective study of 39 pre-operative templates prepared by three different surgeons with different levels of training. The accuracy and reliability measures were evaluated by alpha and kappa coefficients of agreement. The analysis of the effect of coronal deformity on the accuracy of the template was measured by the Spearman’s correlation test. Results Templating was accurate for the femoral component in 28.21% of anterposterior (AP) radiographs and 35.90% of lateral radiographs. Kappa coefficients were respectively 0.111 (95% confidence interval [95%CI]: –0.19 to 0.241) and 0.200 (95%CI: –0.010 to 0.401), indicating poor agreement. Templating accuracy for the tibial component were, respectively, 37.61% and 47.01% for AP and lateral views. Kappa coefficients were 0.186 (95%CI: –0.070 to 0.379) for the AP view and 0.315 (95%CI: –0.199 to 0.431) for the lateral view, showing poor and slight agreement respectively. Considering a margin of error of ±1 sizes, the agreement level improved for all components, particularly for tibia, where agreement levels become very good. The inter-observer agreement was fair for all components, except for the lateral view of the femoral component, whose agreement was good. The Spearman correlation test showed no correlation between accuracy of templating and coronal deformity. Conclusion Pre-operative templating is an unreliable and inaccurate tool. There is no relation between coronal deformity and accuracy of templating.