Reliability of a Caton-Deschamps-derived patella height index for knee arthroplasty

The original Caton-Deschamps index (oCDI) detects functional patella height. It cannot be used in knees with an implanted endoprosthesis. The "modified Caton-Deschamps index" (mCDI) for knee arthroplasty can miss pseudo-patella-infera (PPI), which is common after TKA. A derivate of the oCDI could be a simple analogue to the index published in 1982 using a modified tibial reference point at the anterior proximal point of the inlay, which can indirectly be located on the lateral knee radiograph. It was the aim of this study to determine the intra- and inter-rater agreement of a derived Caton-Deschamps index (dCDI) for knee arthroplasty. We hypothesized that the derived Caton-Deschamps index (dCDI) is a reliable radiological measure for patella height in knee arthroplasty.

Several patella height indices were measured by three independent raters in two passes. The second pass was performed after 6 weeks in random order. Intra- and inter-observer agreements were determined and analyzed using the intraclass correlation coefficient (ICC). For radiographic evaluation, digital lateral radiographs of 150 knees before and after primary TKA were used.

We found high interrater reliability for all analyzed indices. We found the highest agreements for the ISI preop (ICC = 0.914) and postop (ICC = 0.920), respectively. We also found very good intra-rater reliability for the CDI (ICCpreop = 0.954), dCDI (ICCpostop = 0.945), ISI (ICCpreop = 0.960; ICCpostop=0.940) and BPI (ICCpreop = 0.969; ICCpostop = 0.955). Fourteen cases (9.3%) with insignificant PPI were found.

The derived Caton-Deschamps index (dCDI) can easily be used in knee arthroplasty and demonstrated high intra- and interrater agreement, which was similar to other commonly used and established patella height indices.