Background: Recent NICE Guidelines have emphasized the need to have in place by 2007 the capability of offering screening to all women in the first trimester using a combination of maternal age with the ultrasound marker nuchal translucency thickness (NT) and the maternal serum biochemical markers free β-hCG and pregnancy-associated plasma protein-A (PAPP-A). Laboratories will therefore need to consider how to introduce the biochemical component of screening. With the recent launch of these assays on the DPC Immulite 2000 platform, it is appropriate and timely to investigate their clinical and analytical performance on a high throughput immunoassay analyser.
Methods: Within-run and between-day precision was assessed in the normal way. Bias was assessed by comparing samples from normal pregnancies (n = 813) and pregnancies with Down's syndrome (n = 60) run on both the DPC system and our routine Kryptor system. Gestational day-specific medians for each marker were calculated from the unaffected population. Mathematical modelling was used to predict the clinical performance of the two markers.
Results: The within-run coefficient of variation (CV) was around 3.5% and between-day CV was around 6-8% for both assays. Comparison with EQA samples showed a 2% positive bias against the ALTM for free β-hCG and a 21% positive bias for PAPP-A, which reduced to 11% when compared with the Kryptor method mean. When compared with the Kryptor, the DPC free β-hCG showed a significant concentration-related negative bias above concentrations of 50 IU/L in both unaffected and affected pregnancy samples. After conversion to MoM in the Down's syndrome group, the DPC free β-hCG assay showed a MoM-related bias at higher MoM values. For PAPP-A a positive MoM-related bias was also evident. In cases with Down's syndrome, the median free β-hCG MoM was 1.703 for DPC and 1.698 by the Kryptor, which was not significantly different. For PAPP-A, the median MoM was 0.62 for DPC and 0.47 by the Kryptor, which was significantly (P = 0.025) different. This difference, if real, is likely to result in a lower detection of cases with Down's syndrome. Statistical modelling techniques suggest that this would be 58% compared with the expected 67% when combined with maternal age and free β-hCG. When combined with NT, however, this would be reduced to 85% compared with the 90% expected.
Conclusions: While laboratories may find some organizational benefit from running first trimester Down's syndrome screening on such routine high-throughput immunoassay analysers, better clinical performance is likely to be achieved with alternative platforms to the DPC Immulite 2000.
