Comparison Between Bone Ages Assessed By Several Methods

Objective: At present, the commonly used methods of bone ageassessment is G-P Atlas, TW2, TW3, CHN and China-05methods. Differentmethods came from different populations of different regions, they haveregional, ethnic and timing characteristics. In the clinic, we found that thereare great differences between bone ages assessed by different methods. It isnot convenient to communicate that different endocrine centers use differentmethods. Which method is best for children? It is a problem for clinicalendocrine doctors. The bone ages assessed by eight methods were compared totest the applicability of them in contemporary Hebei area girls of different agegroups.Method:Our study included240health girls between4and15years oldreferred to the department of pediatrics endocrinology, nutrition, growthdevelopment at the second hospital of Hebei Medical University, for the freehealth examination in August2011. Girls were considered as normal if theheight and weight were within normal range (mean±2SD).Girls withendocrine disorders or used hormone drugs were excluded from the study.According to the ages,240girls were divided into five groups,4~6years old,6~8years old,8~10years old,10~12years old,12~15years old. The boneages of children were estimated by G-P Atlas, CHN, TW2-RUS, TW2-C,TW2-T, TW3,RUS-CHN, TW3RUS-C methods. The differences betweenbone age (BA) and chronological age (CA) for every method were calculatedand tested by One Sample T Test. The differences between methods within thesame age group were tested by Nonparametric Tests.Spearman’s rankcorrelation was used to evaluate correlation between BA and CA. A p value of<0.05was considered statistically significant. Results:14～6age group, the means of differences between BA estimated byeight methods and CA were-0.307,0.081,0.330,0.333,0.241,-0.822,-0.452,-0.663. In addition to the CHN method (P>0.05), the means of differencesbetween BA and CA had significant differences with0test values (P<0.05).BA estimated by eight methods and CA were positively correlated, thecorrelation coefficient r were0.809,0.782,0.749,0.807,0.806,0.754,0.806,0.754. P values were less than0.01.26～8age group, the means of differences between BA estimated byeight methods and CA were-0.208,-0.003,0.292,0.508,0.367,-0.950,-0.650,-0.742. In addition to the CHN method (P>0.05), the means of differencesbetween BA and CA had significant differences with0test values (P<0.05).BA estimated by eight methods and CA were positively correlated, thecorrelation coefficient r were0.599,0.641,0.492,0.666,0.675,0.465,0.439,0.395. P values were less than0.05.38～10age group, the means of differences between BA estimated byeight methods and CA were0.237,0.459,0.611,0.580,0.194,-0.279,-0.107,-0.160. In addition to the G-P Atlas, RUS-CHN and TW3-C RUS methods(P>0.05), the means of differences between BA and CA had significantdifferences with0test values (P<0.05).Among them, CHN, TW2-R andTW2-C methods had notable significant differences (P<0.01). The differencesamong G-P Atlas,RUS-CHN and TW3-C RUS,tested by Nonparametric Tests,had significant differences (P<0.01). The difference between G-P Atlas andRUS-CHN had significant differences (P<0.01),the difference between G-PAtlas and TW3-C RUS had significant differences (P<0.01), but, thedifference between RUS-CHN and TW3-C RUS had no significantdifferences(P>0.05). BA estimated by eight methods and CA were positivelycorrelated, the correlation coefficient r were0.581,0.465,0.508,0.583,0.572,0.519,0.514,0.534. P values were less than0.01.410～12age group, the means of differences between BA estimated byeight methods and CA were0.070,0.304,0.326,0.043,-0.399,-0.024,-0.104,-0.289.In addition to the G-P Atlas, TW2-C,TW3and RUS-CHN methods (P>0.05), the means of differences between BA and CA had notablesignificant differences with0test values (P<0.01). The differences among G-PAtlas, TW2-C,TW3and RUS-CHN,tested by Nonparametric Tests, had nosignificant differences（P>0.05）. BA estimated by eight methods and CA werepositively correlated, the correlation coefficient r were0.305,0.335,0.402,0.263,0.352,0.393,0.374,0.388. P values were less than0.05.512～15age group, the means of differences between BA estimated byeight methods and CA were0.770,0.700,1.291,-0.124,0.115,0.527,-0.015,-0.348.In addition to TW2-C,TW2-T and RUS-CHN methods (P>0.05), themeans of differences between BA and CA had notable significant differenceswith0test values (P<0.01). The differences among TW2-C,TW2-T andRUS-CHN,tested by Nonparametric Tests, had no significant differences（P>0.05）.6In Each age group, the difference between TW3and TW3-C RUSmethod had notable significant differences (P <0.01).7In addition to the4~6age group (P>0.05), the differences amongTW2-R,TW2-C and TW2-T had significant differences (P<0.05).8BA estimated by eight methods and CA were positively correlated, thecorrelation coefficient r were0.808,0.816,0.806,0.694,0.775,0.835,0.797,0.811. P values were less than0.01.Conclusion:1CHN method is applicable for girls before eight years.2RUS-CHN, TW3-C RUS and RUS-CHN methods are applicable forgirls aged8～10years. The RUS-CHN method is the best.3The G-P Atlas, TW2-C,TW3and RUS-CHN methods are applicable forgirls aged10～12years. They have no difference.4TW2-C,TW2-T and RUS-CHN methods are applicable for girls aged12～15years. They have no difference, but The RUS-CHN method is the best.5RUS-CHN method is applicable for girls after the age of eight.