A retrospective evaluation of second-trimester serum screening for fetal trisomy 18: experience of two laboratories.

Abstract

A retrospective study on screening methods for fetal trisomy 18 has been carried out in two different laboratories using the serum parameters: total human chorionic gonadotropin (hCG), unconjugated oestriol (uE3), and alpha-fetoprotein (AFP) in different combinations and in single marker protocols. Laboratory A (L(A)) utilized a radio-immunoassay to examine 38 fetal trisomy 18 cases and laboratory B (L(B)) utilized an enzyme-immunoassay to examine 33 trisomy 18 cases. As unaffected references the whole routine cohorts of each laboratory were used (L(A): 29 043; L(B): 4264). In both trisomy 18 study groups the median hCG and uE3 multiples of the median (MoM) values were markedly declined (L(A): 0.21 MoM, 0.37 MoM; L(B): 0.31 MoM, 0.44 MoM). Even after exclusion of trisomy 18 cases with combined neural tube or ventral wall defects the medians of AFP MoM values were only moderately declined (L(A): 0.73 MoM; L(B): 0.8 MoM). Receiver-operator characteristic (ROC) curves after multivariate discriminance analysis and single marker evaluation demonstrated that the difference of efficiency between a combination of hCG, uE3 and AFP, and a combination of hCG and uE3 is small but that any of these combinations are more efficient than a combination of hCG and AFP or single marker protocols, respectively. At a risk cut-off generating a false-positive rate of one per cent the most effective marker combination detected 31 of 38 (81.6 per cent) affected pregnancies in L(A) and 25 of 33 (75.8 per cent) in L(B). The differences in sensitivity and specificity seem to be due to the different analytical systems being utilized by the two laboratories.