目的 利用全自动荧光免疫分析仪(GSP)检测先天性甲状腺功能减低症(CH)筛查的指标-促甲状腺素(TSH),评估GSP性能,并建立河南地区新生儿CH在GSP上筛查TSH的切值。方法 1)通过对美国疾病预防与控制中心(CDC) 的滤纸干血片标本的检测,计算GSP的精密度与正确度;检测6个浓度的线性样本计算 GSP的线性; 2)收集用Victor₂D 1420检测的1 273例新生儿样本在GSP上同步检测TSH,通过配对t检验,分析两种检测结果的一致性;3)采集新生儿足跟血,使用GSP促甲状腺素测定试剂盒检测干血片中的TSH,采用百分位数法和ROC曲线法确定TSH切值。结果 1)GSP筛查系统的批内变异系数为4.32%~5.87%;批间变异系数为4.45%~9.08%,均＜1/3总误差(TEa);GSP系统检测的偏移为－7.89%~0.12%,指标的偏移小于1/2TEa;线性标本的检测值和理论值的线性方程为y=0.908x-0.915,相关系数的平方(R²)为0.998;2) GSP 与 Victor₂D 1420检测的1 273例新生儿样本TSH值差异具有统计学意义(t=7.764,P<0.01)。3)百分位数法第99%可信限位为6.1 μU/ml; ROC曲线法临界值为6.45 μU/ml。结论 GSP系统对TSH检测的定量性能可以达到实验室要求,应用于新生儿CH筛查是可行的,但需重新设定切值,切值定在6.1 μU/ml。

Objective To evaluate the feasibility of genetic screening processor(GSP) applied in the screening for congenital hypothyroidism(CH) in newborns by detecting thyroxine stimulating hormone(TSH) in blood, and to establish the cut-off value of TSH in screening for CH in newborns in Henan province.  Methods 1) The precision and accuracy of GSP were calculated by detecting the dired blood spots specimen of Centers for Disease Control and Prevention. The linear sample of 6 concentration gradient specimens was tested to calculate the linearity of GSP. 2) The samples of 1 273 newborns were tested by both Victor₂D 1420 and GSP, and the consistency of the two Methods was analyzed by paired t test. 3) The neonatal heel blood samples were collected and the TSH in the dried-blood spots was detected using the Perkin Elmer GSP test kit. TSH cut-off value was determined by the percentile method and the receiver operating characteristic(ROC) curve method. Results 1) The within-run coefficient of variation of GSP screening system was 4.32%—5.87%. The between-run variation was 4.45%—9.08%, which were less than 1/3 of the total error(TEa). The bias detected by the GSP system was -7.89%—0.12%, less than 1/2 TEa. The linear equation for observed TSH value by expected value was y=0.908x-0.915, and the correlation coefficient(R²) was 0.998. 2) The difference on TSH level in 1 273 neonatal samples detected by GSP and Victor₂D 1420 was statistically significant(t=7.764, P<0.01). 3) The cut-off value of TSH determined by percentile method of 99% confidence interval was 6.1 μU/ml and it was 6.45 μU/ml by the ROC curve method. Conclusion The quantitative performance of the GSP system can meet the laboratory requirements, and it is proved to be a feasible method for neonatal CH screening. However, the cut-off value of TSH needs to be reset to 6.1 μU /ml.