Urinary screening tests for fetal Down syndrome: I. Fresh beta-core fragment

Variable results have been reported using urine beta-core fragment as a marker for fetal Down syndrome. Initial studies by Cuckle et al. (1994) and Canick et al. (1995) indicated that beta-core fragment was an outstanding marker, detecting >80 per cent of Down syndrome cases. Since these reports, widely varying results have been published, indicating between 20 per cent and 66 per cent detection of cases at 5 per cent false-positive rate. The wide variation in the reported data has led to a loss of enthusiasm for this marker as a useful test for Down syndrome screening. Here we report the results of a three-year prospective study in which urine samples were collected daily from women undergoing fetal karyotype analysis for advanced maternal age. Samples were tested within one week of collection and then frozen. We also investigated the likely causes of the variability observed in beta-core fragment data. We collected 1157 urine samples over 955 days. Beta-core fragment levels were measured. A regression line was calculated for the weekly medians of the 1134 control samples and multiples of the control median (MoM) were determined. The median MoM for the controls was 1.0 and the logarithmic standard deviation (log SD) was 0.41. The median MoM for the 23 Down syndrome cases was 5.44 and the log SD was 0.45. Over the study period, 65 per cent of Down syndrome cases exceeded the 95th centile of the control group. The median MoM of control samples and the proportion of Down syndrome cases detected by the test was relatively constant during the study period. The unaffected cases were divided into three equal divisions, corresponding to approximately the first, second and third year of sample collection. No trend was found in the median control MoM values in three sample collection periods (r2=0.04). A similar number of cases exceeded the 95th centile of control samples in the three sample collection periods, 63 per cent, 66 per cent and 66 per cent. Consistent results were indicated during the three years of sample testing. Levels of total oestriol were determined in urine samples and MoM statistics derived. The median oestriol level in Down syndrome cases was 0.59 MoM. Only 12 per cent of cases had MoM levels below the fifth centile. Gaussian models were prepared combining biochemical data and maternal age distribution. While beta-core fragment by itself detected 65 per cent of Down syndrome cases, beta-core fragment modelled with maternal age detected 66 per cent, and modelled with age and total oestriol levels detected 82 per cent of cases at 5 per cent false-positive rate. At the completion of the study, we thawed and reassayed 20 random urine samples (10 control and 10 Down syndrome) collected at different times during the study period. While the control samples (74-1700 ng/ml) had slightly increased values when reassayed (mean value 137 per cent of original prospective value), the Down syndrome samples (360-20,500 ng/ml) all had decreased values when reassayed (mean=53 per cent, t-test, controls versus cases, p = 0.0003). The Down syndrome samples were decreased to between 93 per cent and 12 per cent of the original value. A relationship was identified between the magnitude of the original beta-core fragment value and the change in immunoreactivity when reassayed (r2=0.998). The higher the initial beta-core fragment value the greater the loss of immunoreactivity. We considered the possibility that the beta-core fragment molecules aggregate upon storage in the freezer. We repeated the assay of the 20 samples after treatment with a high salt buffer. Down syndrome samples recovered half of the lost beta-core fragment immunoreactivity (mean increase in beta-core fragment levels 56 per cent, t-test, controls versus cases, p=0.004). We infer that aggregation of beta-core fragment upon storage interferes with beta-core fragment measurements. This may be the cause of the poor beta-core fragment screening performance reported using sto