OBJECTIVE: The aim of this study was to develop a new radiographic measure, the patellar shift index (PSI), for the precise estimation of patellar position relative to the trochlea after a total knee replacement with an unresurfaced patella. METHODS: This study included radiographs of 51 patients suffering from anterior knee pain following total knee arthroplasty. Patellofemoral axial radiographs were analyzed to compare the reliability of the PSI to the classical measures of patellofemoral congruence, the lateral patellar tilt (LPT) and patellar displacement (PD). Intra-rater reliability of the PSI, LPT and PD was estimated using the intraclass correlation coefficient (ICC) and the inter-rater reliability using Krippendorff's alpha (Kα). Agreement proportion of was calculated for the PD. Face validity of the PSI was also tested. RESULTS: The PSI had excellent intra (mean ICC=0.91) and inter-rater (Kα=0.92) reliability, as did LPT (mean ICC=0.96; Kα=0.89). The calculation of PD caused a low level of agreement (47.1%) between evaluators in terms of which images could be measured. The exclusion of these radiographs resulted in a high PD intra (ICC=0.96) and inter-rater (Kα=0.97) reliability. CONCLUSION: The PSI appears to be a reliable and valid measure for patellofemoral congruence in a replaced knee joint with an unresurfaced patella. The superiority of the PSI is the result of its consideration of the geometry and size of the replaced knee joint and independence from radiographic magnification.
OBJECTIVE: The aim of this study was to develop a new radiographic measure, the patellar shift index (PSI), for the precise estimation of patellar position relative to the trochlea after a total knee replacement with an unresurfaced patella. METHODS: This study included radiographs of 51 patients suffering from anterior knee pain following total knee arthroplasty. Patellofemoral axial radiographs were analyzed to compare the reliability of the PSI to the classical measures of patellofemoral congruence, the lateral patellar tilt (LPT) and patellar displacement (PD). Intra-rater reliability of the PSI, LPT and PD was estimated using the intraclass correlation coefficient (ICC) and the inter-rater reliability using Krippendorff's alpha (Kα). Agreement proportion of was calculated for the PD. Face validity of the PSI was also tested. RESULTS: The PSI had excellent intra (mean ICC=0.91) and inter-rater (Kα=0.92) reliability, as did LPT (mean ICC=0.96; Kα=0.89). The calculation of PD caused a low level of agreement (47.1%) between evaluators in terms of which images could be measured. The exclusion of these radiographs resulted in a high PD intra (ICC=0.96) and inter-rater (Kα=0.97) reliability. CONCLUSION: The PSI appears to be a reliable and valid measure for patellofemoral congruence in a replaced knee joint with an unresurfaced patella. The superiority of the PSI is the result of its consideration of the geometry and size of the replaced knee joint and independence from radiographic magnification.