Objective:To establish the reference interval of alpha-fetoprotein(AFP)in Xiangxi by indirect method,and to explore the feasibility of establishing the reference interval by indirect method.Methods:The AFP detection data of 10907 physical examinees in Chinese Medicine Hospital of Xiangxi Tujia and Miao Autonomous Prefertuce were selected.The normality of the data was analyzed by Kolmogorov-Smirnov test,and the abnormal distribution data was transformed into approximate norma distribution by generalized Box-Cox transform.The outliers were eliminated by Turkey method,the reference interval was obtained by nonparametric method,and the 90％confidence interval(CI)of the reference limit was calculated to evaluate the robustness of the reference interval.The reference interval was verified by reference change value(RCV)and 20 small samples.Results:After generalized Box-Cox transform,AFP data showed an approximate normal distribution.123 cases of AFP data were excluded by Turkey method,and finally 10784 cases(6105 males and 4679 females)were included.There was a statistical difference between the sexes in AFP(P＜0.05).The nonparametric method was used to obtain a relatively stable reference interval of AFP(reference limit 90％CI):male≤6.43 ng/mL(6.31～6.51)and female≤6.09 ng/mL(5.95～6.25).The reference interval of AFP for male and female was verified and passed by 20 small samples.The relative deviation between the new reference interval and the reference interval provided by the manufacturer is lower than the reference change value(RCV).Conclusion:Indirect method was used to establish the reference interval of AFP,which accords with the population in Xiangxi.The indirect method is simple,practical and feasible,and can be used as an alternative to establishing reference intervals in laboratories.
NETL
NSTL
万方数据
