Background Various scoring systems document improvement after TKA, but most are associated with a ceiling effect that may fail to distinguish between patients having different levels of knee function after TKA. We therefore developed a new scoring system for patients with higher levels of flexion to eliminate ceiling effects observed with current systems. Questions/purposes The purposes of this study were (1) to determine whether the high-flexion knee score eliminates the ceiling effect, (2) to assess the validity and responsiveness of the high-flexion knee score, and (3) to determine whether the high-flexion knee score can aid in differentiation of the knee status of patients at the ceiling level. Methods We prospectively studied 165 patients with 201 well-functioning knees who had undergone primary TKA. We obtained Knee Society scores, WOMAC scores, Feller scores, SF-36 scores, and high-flexion knee scores for all patients. The high-flexion knee score includes items that reflect knee function in the high functional range, such as sitting on or rising from the floor, squatting, or kneeling. We determined the ceiling effects and score distributions of various scoring systems. We performed a convergent validity test of the high-flexion knee score by correlation analysis with these various scoring systems. Responsiveness of the high-flexion knee score was assessed by correlation analysis of changes in various scoring systems. To determine whether the high-flexion knee score can aid in differentiation of knee status of patients at the ceiling level, relative responsiveness of the various scores in the ceiling versus below the ceiling range was determined. Results The high-flexion knee score showed no ceiling effect, whereas the other systems did. Addition of the highflexion knee score to the other scoring systems eliminated these ceiling effects and resulted in more normalized score distributions. The high-flexion knee score correlated (r = À0.77) with WOMAC in postoperative scores, and it also correlated with the changes in WOMAC (r = À0.69) and SF-36 physical functioning (r = 0.62).