Urine is a new medium for Down syndrome testing. In an effort to determine the best type of human chorionic gonadotropin (hCG)-related immunoassay for urine testing, we examined 14 Down syndrome and 91 unaffected pregnancy urine samples with 12 established assays. The assays included (a) those that detect hCG beta-core fragment only; (b) those that detect beta-core fragment with less than 18 per cent free beta-subunit cross-reactivity; (c) that which equally detects free beta-subunit and beta-core fragment; and (d) those that detect hCG, free beta-subunit, or combinations thereof. The seven type a and b assays had the highest sensitivity for Down syndrome. The median MOM for Down syndrome was 5.93 (range 4.73-7.53). At a 10 per cent false-positive rate, the median observed detection rate was 93 per cent (range 79-100 per cent) and the median predicted detection rate was 85 per cent (range 69-96 per cent). The assays that did not mainly detect beta-core fragment (types c and d) had poorer screening performance. The median MOM for Down syndrome was 2.70 (range 2.16-3.63 MOM). At a 10 per cent false-positive rate, the median observed detection rate was 50 per cent (range 36-64 per cent) and the median predicted detection rate was 37 per cent (range 21-62 per cent). We infer that the assays that only detect beta-core fragment, or beta-core fragment with minor free beta-subunit cross-reactivity (types a and b), are the better urine-based tests for Down syndrome screening.