Reproducibility of the Greulich & Pyle method for determining the bone age of a Brazilian population sample

Introduction: this study aimed to evaluate the reproducibility of the widely known Greulich & Pyle (G&P) method to determine the bone age of a Brazilian population sample. Methodologies: two trained evaluators evaluated 302 radiographs (153 males and 149 females) on a 17-inch liquid crystal display with 1,280x1,024 pixels resolution and determined the bone age of the participants based on the G&P atlas parameters. Data was analyzed at a level of significance of 5% (Cronbach’s alpha, Fleiss’ Kappa index, Student’s t-test, and Kruskal-Wallis). Results: the overall percentage of agreements between the chronological age of the participants and the bone age determined by evaluators 1 and 2 based on G&P atlas parameters were 12.91% and 12.25%, respectively. The agreements for males were 18.30% and 15.69%, and for females were 7.38% and 8.72% for evaluators 1 and 2, respectively. The determined mean bone age was significantly higher than the mean chronological age (p<.001). Comparisons between overall bone age and chronological age at all 6-month intervals were found significantly different (p<.05). The comparison between genders revealed that the bone age mean of males was significantly higher than females only at the 9.6-year chronological age interval (p<.05). Conclusion: although the bone age of both genders can be determined by using the G&P method, which is based on the chronological age of an American population sample, it seems not an appropriate parameter to be used in a Brazilian population since a significantly higher bone development was observed.