RESUMO | O objetivo deste estudo foi verificar a validade preditiva das equações de referência para predição da força de preensão manual (FPM) em homens brasileiros de meia idade e idosos. Foram avaliados 60 homens com idade de 58, 7±8, 6 anos (50-84 anos), estatura 168, 7±8, 0 cm, massa corporal 67,9±13, 0 kg e índice de massa corporal 23, 7±4, 0 kg/m 2 , utilizando equipamento 'padrão-ouro' na avaliação da FPM e seguindo todas as recomendações metodológicas preconizadas pela American Society of Hand Therapists (ASHT). Os resultados do diagrama de Bland-Altman para mão dominante (MD) apresentam um erro médio (viés) de -8, 4% (IC95% -51, 6-34, 7). Já o diagrama de Bland-Altman para mão não dominante (MND) apresenta um erro médio (viés) de 1, 4% (IC95% -47, 1-49,8). O cálculo de Cronbach's alpha para a MD foi de 0, 69 e 0,59 para MND. O coeficiente de correlação intraclasse para MD foi 0,52 (IC95% 31-68) e de 0, 42 (IC95% 20-60) para a MND. Concluímos que as equações para predição analisadas neste estudo, para uma população de homens brasileiros de meia idade e idosos, apresentou baixa validade preditiva, o que pode gerar erros de interpretação dos resultados.Descritores | estudos de validação; força da mão; medidas; dinamômetro de força muscular.ABSTRACT | The objective of this study was to verify the predictive validity of reference equations for hand grip strength (HGS) in middle-aged and elderly Brazilian men. We evaluated 60 men with age of 58.7±8.6 years, height 168.7±8.0 cm, body mass 67.9±13.0 kg, and body mass index 23.7±4.0 kg/m 2 , utilizing the gold standard equipment for evaluation of HGS and following all methodological recommendations of the American Society of Hand Therapists (ASTH). The results of the BlandAltman diagram for the dominant hand (DH) had a mean error (bias) of -8.4% (95%CI -51.6-34.7). The BlandAltman diagram for the non-dominant hand (NDH) had a mean error (bias) of 1.4% (95%CI -47.1-49.8). The calculation of Cronbach's alpha was 0.69 for the DH and 0.59 for the NDH. The intraclass correlation coefficient was 0.52 (95%CI 31-68) for the for the NDH. We therefore conclude that the reference equations for HGS evaluated in this study has low predictive validity for a population of Brazilian men middleaged and elderly, and may lead to a misinterpretation of the results.

Introduction Muscle fatigue is a phenomenon associated with physical work. It is common in endurance sports, physical fitness tests and daily activities. Some tests can be directly affected by the effect of peripheral muscle fatigue, including the handgrip strength (HGS) test, which is considered baseline measure for assessing the functionality of the hand. Objectives a) verify the effect of peripheral muscle fatigue (between trials) during the testing of HGS, with a 60-second recovery interval; b) to analyze whether there is a difference in considering the mean value obtained in three trials or the best result as the final result. Materials and methods The final sample comprised 1,279 men. We followed the standard methodology and used a hydraulic hand dynamometer. Results There were statistically significant differences (P < 0.05) in the dominant hand among all the trials: first (46.5 ± 8.6 Kgf), second (46.4 ± 8.5 Kgf) and third (46.1 ± 8.6 Kgf); and also in the non-dominant hand: first (44.9 ± 8.4 Kgf), second (44.5 ± 8.3 Kgf) and third (44.0 ± 8.3 Kgf). We also found statistically significant differences (P < 0.05) between the two ways of considering the final result. For the dominant hand, the mean of the three attempts was 46.3 ± 8.3 Kgf and the best result was 48.1 ± 8.5 Kgf, whereas for the non-dominant hand, these results were 44.5 ± 8.2 Kgf and 46.0 ± 8.2 kgf, respectively. Conclusion Peripheral muscle fatigue directly interferes in the final result. A significant reduction in strength levels occurs in course of the assessment. The best result is frequently obtained at the first trial, which indicates that the highest value obtained should be considered as the final result.