Are modern knee outcomes scores appropriate for evaluating anterior knee pain and symptoms after total knee arthroplasty?

BACKGROUND: Up to 45% of patients who undergo primary total knee arthroplasty (TKA) with contemporary implants have residual anterior knee pain. While a specific evaluation of anterior knee symptoms is mandatory, little is known about the capability of patellofemoral scores to be used individually. This study aimed to assess the distribution of patellofemoral scores after TKA from a uniform cohort and to investigate their external validity and ability to detect anterior knee symptoms using floor and ceiling effects. HYPOTHESIS: Patellofemoral scores have high construct validity and fewer floor/ceiling effects than general knee scores. METHODS: We prospectively included 113 consecutive patients who underwent primary TKA for primary osteoarthritis at a single University Hospital. Clinical outcomes included patellofemoral scores (HSS Patella, Kujala and Lille scores) and general knee scores (KOOS and new KSS) at 1-year follow-up. The floor and ceiling effects were considered as significant when greater than 15%. These were determined for each score individually and for composite scores (combination of patellofemoral scores and new KSS). The construct validity of each score and their ability to detect anterior knee pain was evaluated. RESULTS: Patellofemoral scores showed no floor effect but a significant ceiling effect (from 25% to 65%). This ceiling effect decreased when composite scores were used. The convergent validity test showed strong correlation between patellofemoral scores (from 0.741 to 0.819, p<0.00001) and a better ability to discriminate anterior knee pain than the general knee scores. CONCLUSIONS: Patellofemoral scores showed no floor effect and a very good construct validity for anterior knee pain after TKA. However, studies aiming to monitor anterior knee symptoms after TKA should combine scoring systems to included patellofemoral-related items rather than use patellofemoral scores alone due their ceiling effects. LEVEL OF EVIDENCE: III; prospective study.