RESUMO OBJETIVO: Comparar os métodos mais usados de medida da altura patelar, com o método do ângulo platô-patela. MÉTODOS: Foi feito um estudo transversal no qual foram avaliadas radiografias em perfil do joelho, com os três métodos já consagrados pela literatura, o Insall-Salvati (IS), o Blackburne-Peel (BP) e o Caton-Deschamps (CD) e comparando-as com o ângulo platô-patela (APP). Foram incluídos na amostra 196 seis pacientes, aleatoriamente selecionados. RESULTADOS: Inicialmente os dados foram submetidos a uma avaliação pelo teste do qui-quadrado. A análise foi positiva com p < 0,0001. Fizemos comparações entre os métodos tradicionais com a medida do APP com o uso do teste exato de Fisher. Quando comparamos o índice de IS com o APP, não encontramos diferenças estatisticamente significativas em relação à proporção de casos alterados entre os dois grupos. Os métodos tradicionais foram comparados com a medida do APP quanto à proporção de casos de patela alta e baixa pelo teste exato de Fisher. A análise demonstrou que o APP identificou menos casos de patela alta do que os métodos de IS, BP e CD, mas identificou mais casos de patela baixa. Quando comparados os pares, verificamos que os índices de IS e CD foram capazes de identificar mais casos de patela alta que o APP. Em relação aos casos de patela baixa, o APP foi capaz de identificar mais casos que os outros três métodos. CONCLUSÃO: O ângulo platô-patela observou mais pacientes com patela baixa em comparação com os métodos clássicos e resultados discrepantes com os outros índices estudados.

ABSTRACT OBJECTIVE: The methods most used for patellar height measurement were compared with the plateau-patella angle method. METHODS: A cross-sectional study was conducted, in which lateral-view radiographs of the knee were evaluated using the three methods already established in the literature: Insall-Salvati (IS), Blackburne-Peel (BP) and Caton-Deschamps (CD). These were compared with the plateau-patella angle method. One hundred and ninety-six randomly selected patients were included in the sample. RESULTS: The data were initially evaluated using the chi-square test. This analysis was deemed to be positive with p < 0.0001. We compared the traditional methods with the plateau-patella angle measurement, using Fisher's exact test. In comparing the IS index with the plateau-patella angle, we did not find any statistically significant differences in relation to the proportion of altered cases between the two groups. The traditional methods were compared with the plateau-patella angle with regard to the proportions of cases of high and low patella, by means of Fisher's exact test. This analysis showed that the plateau-patella angle identified fewer cases of high patella than did the IS, BP and CD methods, but more cases of low patella. In comparing pairs, we found that the IS and CD indices were capable of identifying more cases of high patella than was the plateau-patella angle. In relation to the cases of low patella, the plateau-patella angle was capable of identifying more cases than were the other three methods. CONCLUSIONS: The plateau-patella angle found more patients with low patella than did the classical methods and showed results that diverged from those of the other indices studied.

Objective: To assess the reproducibility of Tronzo and AO/ASIF classifications for transtrochanteric fractures, in order to determine the most appropriate classification for clinical application, and to evaluate the influence of the level of experience of the observers in the agreement between evaluations.Methods: We selected 30 radiographic images of transtrochanteric fractures of the femur, which were presented to two groups of observers, one formed by expert physicians and the other by resident physicians.Results: When evaluated together, Tronzo classification obtained a Kappa value of 0.44. The same classification assessed by the expert group obtained a value of 0.46, while the group of residents' value was 0.44. Evaluating the AO/ASIF classification of the complete pool analysis the value found was 0.42. For the same classification, analyzed by the expert group, obtained a value of 0.41, and by the group of residents, the Kappa value achieved was 0.42. However, when analyzed in its simplified form, the AO/ASIF classification obtained Kappa values of 0.70 (pooled analysis), 0.68 (experts) and 0.72 (residents), considered concurrent.Conclusion: The AO/ASIF simplified classification showed substantial reproducibility and is, therefore, recommended as the most suitable for clinical application. The level of experience of the observers did not influence significantly the agreement between evaluations. Level of Evidence III, Diagnostic Study - Investigating a Diagnostic Test.