Elevated maternal urine level of β-core fragment of human chorionic gonadotropin versus serum triple test in the second-trimester detection of Down syndrome

Objective: This study was undertaken to compare the Down syndrome screening efficiency of elevated maternal urine level of the β-core fragment of human chorionic gonadotropin with that of the traditional serum triple test. Study Design: Urinary β-core fragment and serum analyte levels were measured prospectively in women with singleton pregnancies who were undergoing second-trimester genetic amniocentesis. Urinary analyte levels were measured within a week of specimen collection. In some cases only α-fetoprotein was measured initially and human chorionic gonadotropin and unconjugated estriol levels were subsequently determined from the stored serum specimens. The Down syndrome screening efficiency of urinary concentration of β-core fragment plus maternal age was compared with that of the traditional triple test. Receiver operating characteristic curves were generated for each algorithm and the areas under the curves were compared to determine which algorithm was superior. Results: There were a total of 926 study patients, of whom 21 (2.3%) carried fetuses with Down syndrome. The mean (±SD) gestations at amniocentesis were 16.6 ± 1.5 weeks for the fetuses without Down syndrome and 17.7 ± 2.3 for the fetuses with Down syndrome. A total of 539 women (4 of whom carried fetuses with Down syndrome) had serum α-fetoprotein alone measured initially. Urinary concentration of β-core fragment had a 61.9% detection rate with a 4.9% false-positive rate for Down syndrome, whereas the values for the triple screen were 57.1% and 11.2%, respectively. The areas under the receiver-operating characteristic curves were 0.8744 for elevated urinary β-core fragment level and 0.7504 for the triple screen ( P = .1116). When the false-positive rate was fixed at an ideal threshold value (≤5%) the urine test was superior (area under the curve, 0.0212 vs 0.0133, P < .05). Similarly, when we considered only cases in which the complete triple screen was performed prospectively (17 fetuses with Down syndrome and 431 fetuses without Down syndrome), the urine test was significantly better (area under the curve, 0.873 vs 0.624, P = .012). Conclusion: In this first reported direct comparison we consistently observed higher sensitivity values for screening with urinary levels of β-core fragment than for serum triple screen, suggesting an equivalent or superior Down syndrome screening performance for the urinary analyte. It is important that freezing and prolonged urine storage before testing b