PURPOSE OF THE STUDY: Numerous scoring systems have been described for the evaluation of total knee arthroplasty. The aim of this study was to analyze the reality of the results and the differences between clinical evaluation scores. MATERIAL AND METHOD: We calculated the mean overall scores (Hungerford, Laskin, Hospital Special Surgery and Mansat) in a series of 89 PCA total knee arthroplasties with a mean follow-up of 4.5 years. We compared postoperative score values considered as borderline between poor and good results with survival curves. RESULTS: The mean values of the postoperative scores were very similar and showed no significant differences: Hungerford: 83 pts; Laskin: 80 pts; HSS: 80 pts; and Mansat: 86 pts. The comparison between the four survival curves was statistically significant: the Hungerford curve was the most optimistic (62 per cent success at 8 years) when compared to the Laskin score (16 per cent success at 8 years). DISCUSSION: The scores do not express the same reality. For mean values close or identical to the scores, the quality of the results may be different. Our analytic method reveals considerable differences explained by the juxtaposition of clinical and functional parameters in the scores. The Hungerford and Mansat scores are principally based on clinical criteria while the Laskin score is more functionally orientated. The HSS score is more even handed. CONCLUSION: It is incorrect to appreciate the value of survival curves if the scores are based on functional criteria as these parameters will naturally deteriorate over time and therefore these scores do not reflect the real quality of the arthroplasty. The Knee Society Score is probably the most interesting system since functional and clinical parameters are separated.