Reliability of a modified Berger protocol for TKA component rotation using an anatomical tibial component

Malrotation in total knee prostheses can cause persistent pain and dissatisfaction. While the Berger protocol is the standard for assessing component rotation, it is unsuitable for anatomically designed tibial components. This study evaluates the inter- and intra-observer reliability of a modified method for anatomical prostheses and its clinical applicability.

In this prospective study, 500 patients underwent postoperative CT scans. Femoral rotation was measured as the angle between the transepicondylar axis and posterior condylar line. Tibial rotation was defined by the angle between a perpendicular axis on the tibial plateau and a line through the tuberosity. Inter- and intra-observer reliability was assessed using the intraclass correlation coefficient (ICC) with a two-way mixed effect model with single measurement/observer and absolute agreement.To determine the clinical applicability, we calculated the percentage of agreement per degree of deviation between observers.

Inter-observer reliability was good for both femoral (ICC 0.752, 95 % CI 0.71-0.79) and tibial (ICC 0.866, 95 % CI 0.84-0.89) measurements. Femoral measurements differed less than 2° in 91.4 % of cases, versus 52.2 % for tibial. Intra-observer reliability was found excellent: femoral (observer 1: ICC 0.941 (95 % CI 0.929-0.950); observer 2: ICC 0.943 (95 % 0.907-0.954)) and tibial (observer 1: ICC 0.941 (95 % CI 0.921-0.956); observer 2: ICC 0.972 (95 % CI 0.960-0.980).

The modified Berger protocol shows high reliability for anatomical tibial components. However, the percentage agreement for the tibial component between and within observers was low and therefore might influence the applicability.