Abstract：【Objective】To evaluate and adjust the cut-off value of phenylalanine (Phe) in neonatal phenylketonuria screening in Shenzhen by percentile method and ROC curve analysis.【Methods】A total of 559,496 newborns delivered in various hospitals in Shenzhen from January 2018 to December 2020 were selected. The Phe level was detected by fluorescence phenylalanine dehydrogenase method, the current Phe cut-off value was evaluated, and the best cut-off value of Phe level in Shenzhen was established. 【Results】 The Phe level of 559,496 newborns showed a skewed distribution, with a median of 1.06 mg/dL, a minimum of 0.01 mg/dL, a maximum of 24.55 mg/dL, a 95% quantile of 1.69 mg/dL, a 99% quantile of 1.96 mg/dL, a 99.5% quantile of 2.18 mg/dL and a 99.9% quantile of 2.73 mg/dL. 3,923 cases were reexamined with Phe level ≥2 mg/dL, including 3267 cases of screening positive recall. A total of 33 cases were finally diagnosed according to the diagnostic criteria. The sensitivity and specificity of 3,267 newborns at different cut-off values were finally adjusted from the current cut-off value of 2.0 mg/dL to 2.1 mg/dL; The area under ROC curve was 0.960, the standard error was 0.027, and the 95% confidence interval was 0.906~1.000, which was statistically significant (P<0.01). Finally, the current cut-off value of 2.0 mg/dL was adjusted to 2.1 mg/dL. 【Conclusion 】Combined with percentile method and ROC analysis, it is more appropriate to adjust the best cut-off value of Phe for neonatal PKU screening in Shenzhen to 2.1 mg/dL.
[1] 赵正言,顾学范.新生儿遗传代谢病筛查[M].第2版.北京:人民卫生出版社,2015:9.
[2] 赵德华,贾晨路,吕书博,等.ROC曲线在河南省新生儿苯丙酮尿症筛查阳性切值中的应用[J].中国妇幼保健,2016,31(9):1905-1907.
[3] 章晓燕, 王薇, 赵海建,等. 我国2015年新生儿遗传代谢病筛查指标切值的调查[J].临床检验杂志, 2016, 34(9):706-709.
[4] 孙振球.医学统计学[M].北京:人民卫生出版社,2002:9-19.
[5] 中华医学会医学遗传学分会遗传病临床实践指南撰写组.苯丙酮尿症的临床实践指南[J].中华医学遗传学杂志,2020,37(3):226-234.
[6] 顾学范,王治国.中国580万新生儿苯丙酮尿症和先天性甲状腺功能减低症的筛查[J].中华预防医学杂志,2004,38(2):99-102.[7] MARCO Z, GIORGIO Z, CLAUDIO M, et al. Neonatal birth parameters of positive newborns at PKU screening as predictors of false-positive and positive results at recall-testing[J].J Med Screen,2003,10(4):181-183.
[8] 沈隽霏,宋斌斌,潘柏申.间接法建立生物参考区间[J].检验医学,2015,30(4):391-396.
[9] 国家卫生健康委员会临床检验中心新生儿遗传代谢病筛查室间质评委员会.新生儿遗传代谢病筛查指标切值建立方法专家共识[J].中国实用儿科杂志,2019,34(11):881-884.
[10] 陈卫中,倪宗瓒,潘晓平,等. 用ROC曲线确定最佳临界点和可疑值范围[J].现代预防医学,2005,32(7):729-731.
2021, Vol. 38 
