Discriminant analysis using in congenital adrenal hyperplasia neonatal screening

BAO Pei-zhong,YE Jun,HAN Lian-shu,QIU Wen-juan,ZHANG Hui-wen,ZHOU Jian-de,GU Xue-fan
2013-01-01
Abstract:Objective To evaluate the discriminant method using for the neonatal congenital adrenal hyperplasia(CAH) in neonatal screening. Methods The retrospective CAH screening data during 2008-2011 from our center were investigated. The source data was separated into 2 group based on the traditional CAH screening indicators 17-OHP cut-off 30 nmol/L. 11 130 cases with 17-OHP value<30 nmol/L were selected randomly as the normal group, 398 cases including congenital adrenal hyperplasia patients(CAH) with 17-OHP value ≥30 nmol/L were set as the positive group. A discriminate equation and a new positive cutoff which involving the neonatal gestational age, weight, and 17-OHP value were set up by using of the unstandardized Fisher analysis program. All positive samples including 24 CAH patients and 381 false positive cases which were validated by the traditional 17-OHP cut off were re-evaluated with new set criterion. Results A discriminant equation base on the source data were established as Z=(-0.019·x1)+(-0.00012·x2)+0.055·x3 +0.471, and the new positive cut off was set as 1.314. The data of 24 real CAH patients was verified with new set discriminating formula, and the result was confirmed as 100%(24/24) agreed as its previous diagnosis. 27%(156/573) false positive cases were change into negative after the data was re-evaluated by new method. Conclusion The multi-factor discrimant analysis for neonatal CAH screening will expect to significantly reduce the false positive rate, therefore it show more reasonable and feasible in neonatal screening.
What problem does this paper attempt to address?