68

R e v i e w s MOLECULAR MEDICINE TODAY, FEBRUARY 1999

1357-4310/99/$ - see front matter © 1999 Elsevier Science. All rights reserved. PII: S1357-4310(98)01396-3

DOWN syndrome is the most common, serious, chromosomal ab-normality identified after birth that affects ~1 in 700 newborns and isone of the leading causes of mental retardation. It is caused by thepresence of an extra copy of chromosome 21, or a portion of that

chromosome, in each cell of an affected individual. Most commonly,the mechanism for this error is the inability of the two copies ofchromosome 21 to separate during meiosis (nondysjunction); thisoccurs most frequently in the maternally derived gamete, the oocyte.When a haploid oocyte with two copies of chromosome 21 is ferti-lized by a normal haploid sperm, trisomy 21 or Down syndrome oc-curs. In females, this gives rise to a karyotype of 47 XX (includingan additional chromosome 21) and 47 XY in males.

Maternal nondysjunction increases in frequency with advancingmaternal age so that, for example, a 20 year old has a ~1 in 1500 riskof having an affected baby but a 30 year old has a 1 in 900 risk. Bythe time a woman is 40 years old, her risk of having a baby withDown syndrome has increased to 1 in 100. During the late 1960s andearly 1970s, an amniocentesis test became available in the secondtrimester of pregnancy in which a karyotype analysis of the cells ofthe amniotic fluid was performed. So-called genetic amniocentesisis highly accurate (it has almost no errors in detection) but it is an in-vasive procedure that can result in miscarriage. The most carefullydocumented estimated risk of miscarriage caused by amniocentesis isthe loss of one pregnancy in 100 procedures1. It was the associationof maternal age with risk that led to the first prenatal screening testfor Down syndrome, because amniocentesis could be justified inpregnant women who were at a high risk on the basis of their age.

With maternal age as the screening test, ~25% of all Down syn-drome pregnancies could be identified by offering amniocentesis tothe oldest 5% of the pregnant population (in 1980, the oldest 5% was>35 years). This meant, however, that 75% of babies with Downsyndrome were still undiagnosed because they were born to mothersunder the age of 35, women who were considered at too low a risk tobe offered the invasive, risky diagnostic procedure of amniocentesis.

Maternal age remained the only method of screening pregnantwomen for risk of Down syndrome until 1984. The ability to identify

Urinary analyte screening: anoninvasive detection method forDown syndrome?

Jacob A. Canick, Leonard H. Kellner, Laurence A. Cole and Howard S. Cuckle

Prenatal screening for Down syndrome using maternal serum markers achieves detection rates of 60–80%with a 5% false positive rate. Improvement in the accuracy of screening, as well as its ease and safety, willincrease the use of such tests. The most effective of the current serum markers is human chorionicgonadotropin (hCG). Studies on b core fragment (bCF), the major urinary metabolite of hCG, haveindicated that screening with bCF and other markers measured in maternal urine might improve thedetection of Down syndrome and provide a less expensive and simpler test. However, recent results havebeen unusually variable. Although it has great potential, the true clinical value of maternal urine screeningto detect Down syndrome still remains to be determined.

Jacob A. Canick* PhDProfessor

Dept of Pathology and Laboratory Medicine, Women and InfantsHospital, Brown University School of Medicine, Providence, RI, USA.

Tel: 11 401 453 7654Fax: 11 401 453 7547

*e-mail: jcanick@wihri.org

Leonard H. Kellner MSInstructor

Dept of Obstetrics and Gynecology, Montefiore Medical Center,Albert Einstein College of Medicine, Bronx, NY, USA.

Laurence A. Cole PhDAssociate Professor

Dept of Obstetrics and Gynecology, Yale University School ofMedicine, New Haven, CT, USA.

Howard S. Cuckle DPhilProfessor

Centre for Reproduction, Growth and Development, School ofMedicine, University of Leeds, Leeds, UK.

high-risk individuals among younger preg-nant women began with the discovery thatmaternal serum levels of a-fetoprotein(AFP), a fetal liver product used in prenatalscreening for neural tube defects, tended tobe low in Down syndrome pregnancy2,3. Bythe mid 1980s, risk assessment using mater-nal serum AFP, in combination with maternalage, resulted in the identification of 25% ofthe Down syndrome pregnancies in womenunder the age of 35. Following this discov-ery, other products of the fetoplacental unitthat are measurable in maternal serum wereexamined. Currently, the most informativeserum markers used in screening are humanchorionic gonadotropin (hCG)4, the free bsubunit of hCG (free b-hCG)5 and inhibin A(inhA)6, all of which tend to be elevated, andunconjugated estriol (uE3)7, which tends tobe low in Down syndrome pregnancy (re-viewed in Ref. 8).

When applied to all pregnant women, a two-marker combination(AFP plus hCG or free b-hCG) results in the identification of 59% ofDown syndrome pregnancies, a three-marker combination (that in-cludes uE3) results in the identification of 69% of Down syndromepregnancies, and a four-marker combination (that includes inhA) re-sults in the identification of 76% of Down syndrome pregnancies inthe 5% of the screened population that has the highest calculatedrisk8,9. The current standards of care dictate that some form of mul-tiple marker screening, be it of two, three or four markers, should beoffered to pregnant women.

Of the analytes currently used, hCG, free b-hCG and uE3 areknown to be present in maternal urine, as well as in maternal serum,thus indicating that urine could be used to screen for Down syndrome(see Table 1). Recent studies have shown that the levels of these ana-lytes and their metabolic products might, in fact, be markedly abnor-mal in maternal urine samples from Down syndrome pregnancies.The purpose of this review is to examine the potential use of maternalurine analytes in prenatal screening for Down syndrome.

Potential benefits of urine screeningAs non-invasive and as simple as serum screening might be, there arepractical implications that apply to the collecting and handling ofblood samples that perhaps make the availability of such screeningmethods sub-optimal. Such issues include the need for phlebotomyservices, the requirement for patients to attend a clinic to have ablood sample taken, the regulations that affect the transport of sam-ples and the inherent costs of this procedure.

If screening could be accomplished in a single sample of maternalurine, as opposed to blood, then the issues of specimen collection,transport and cost could be eliminated or reduced. Urine could becollected at home, removing the need for patients to make an ap-pointment and travel to blood-drawing centers, and for screeningprograms to finance the cost of blood sampling. It would also lessenthe exposure of technical personnel to potentially infectious diseases.Ultimately, the use of urine could substantially increase the availabil-ity and use of prenatal screening.

Although urine-based screening might have several benefits, itspotential limitations must also be considered. First, urine samples are

less uniform than serum because of large fluctuations in volume andsolute concentration that depend on the time of day that the sample istaken and the dietary status of the individual. Such fluctuations canbe controlled by using standardized collection times and normaliz-ation methods, but these must be examined critically. Second, thecollection of urine samples is not usually done under sterile condi-tions. Therefore, anti-microbial agents would have to be tested andbe introduced into the collection device. Third, maternal serum willremain a necessary part of prenatal screening for neural tube defectsbecause AFP, the analyte used for neural-tube-defect screening, is notreadily measurable in urine using the current assay methodology, if itis measurable at all. Nevertheless, specialized centers that routinelyscan for anomalies at midtrimester can detect most neural tube de-fects without using biochemical assays. (See Box 1 for the advan-tages and disadvantages of urine versus serum screening in the pre-natal diagnosis of Down syndrome.)

hCG synthesis and metabolismhCG is a glycoprotein hormone that comprises two subunits, a andb. It is synthesized by the syncytiotrophoblast layer of the placentalvilli and is secreted preferentially into the maternal circulation. hCGhas the important function of stimulating the corpus luteum during

69

R e v i e w sMOLECULAR MEDICINE TODAY, FEBRUARY 1999

Table 1. Median change in biochemical marker levels in maternalserum and maternal urine from Down syndrome pregnancy as

compared with unaffected pregnancya

Median change in

Marker Maternal serumb Maternal urine

hCG 2.13 2.33

Free b-hCG 2.23 2.33

bCF N.p. 4.13

Estriol 0.73c 0.63d

aAbbreviations: bCF, b-core fragment; hCG, human chorionic gonadotropin; N.p., not present.bMaternal serum medians are taken from metaanalyses of the world literature8.cUnconjugated estriol.dTotal estriol.

Box 1. Advantages and disadvantages of urineversus serum for Down syndrome screening

Advantages of serum screening

• Samples are sterile

• Samples are uniform

• Can simultaneously screen for open neural tube defects

Advantages of urine screening

• Collection is cheap

• Collection is simple

• Lower risk to laboratory staff

• More comfortable for patient

the first ten weeks of pregnancy to produce sufficient progesterone tokeep the uterine environment optimal for embryonic development. Italso is known to stimulate the fetal testes to produce testosterone forthe normal masculinization of a male fetus. The a and b subunits areseparate gene products and have molecular weights of 14.5 kDa and22.2 kDa, respectively. The regulation of the synthesis of hCG ap-pears to be complex, with many autocrine and paracrine factorsdemonstrated to have positive and negative effects in experiments invitro10. hCG is present in the maternal circulation not only in its intactform but also as its individual free subunits and in various metabolicand degraded forms.

The metabolism of hCG is believed to follow a pathway that be-gins with the hydrolysis of specific peptide bonds in the b subunit of

the hCG dimer (the so-called ‘nicking’ of hCG; see Fig. 1). NickedhCG then rapidly separates into its component subunits. The b sub-unit can undergo further substantial proteolysis, which leads to therelease of a C-terminal peptide and a final metabolic form called theb-core fragment (bCF), which comprises two disulfide-bonded corepeptides of ~10 kDa. bCF is the major hCG metabolite in urine, al-though all forms of hCG and its various metabolites are measurablein urine11,12.

In addition to the hCG metabolites described above, a percentageof hCG in the maternal circulation is ‘hyperglycosylated’, in that ithas extra branching on certain carbohydrate substituents of the a andb subunits. For example, ~5% of circulating hCG is hyperglycosyl-ated during the second trimester. Hyperglycosylated hCG is be-

lieved to break down more rapidly than nor-mally glycosylated hCG, and it might bepresent in higher concentrations in abnormalthan in normal pregnancy13. The reason forthe more rapid breakdown could be relatedto the loosening of the usual hydrogen bondsbetween the a and b subunits of hCG; thisloosening is caused by the extra carbohy-drate residues in hyperglycosylated hCG.

Estriol synthesis and metabolismEstriol is a steroid hormone that has weakestrogenic (female sex hormone) activity. Itis peculiar to pregnancy and increases inconcentration in maternal blood throughoutgestation, beginning at about eight weeksafter the last menstrual period. Estriol is syn-thesized by the placenta from steroid precur-sors that are supplied by the fetus (Fig. 2).Dehydroepiandrosterone (DHEA) sulfate,which is synthesized by the fetal adrenal cor-tex from cholesterol, is 16a-hydroxylated inthe fetal liver to 16a-hydroxyDHEA sulfate.This then passes through the umbilical circu-lation to the placenta where it is metabolizedand finally aromatized to estriol. Estriol inits nascent, unconjugated form (uE3) is se-creted preferentially into the maternal blood,where it circulates with a very short half-lifeof ,30 min. Unlike estradiol, uE3 does notcirculate bound to sex hormone-bindingglobulin. UE3 is rapidly metabolized to moresoluble estriol conjugates, sulfates and glu-curonides, which are efficiently filtered intothe urine.

In urine, the measurement of total estriol(tE3; unconjugated plus conjugated) reflects,to a large degree, the level of unconjugatedestriol in serum, although other estrogens besides estriol contribute to the final total estriol pool in urine.

Urinary metabolites of hCG and estriol inDown syndrome pregnancyThe analysis of hCG metabolites in maternalurine as possible markers of Down syndrome

70

R e v i e w s MOLECULAR MEDICINE TODAY, FEBRUARY 1999

Figure 1. The fast and slow pathways of human chorionic gonadotropin (hCG) dissociation and metab-olism, which leads to the formation of b-core fragment. The b subunit and its metabolites are shown in redand the a subunit in blue. Black lines within the subunits indicate disulfide bonds.

Human chorionicgonadotropin (hCG)

β subunit α subunit

Nicked hCGNon-nickedβ subunit

Free α subunit

Nicked freeβ subunit Free α subunit

β-core fragment

Degradation

Rapid dissociationNicking

Slow dissociation Nicking

pregnancy began with the studies of Cuckleand colleagues14. They speculated that, be-cause the cascade of hCG metabolism is be-lieved to begin with the nicking of the hCGb subunit and because there appeared to bemore nicking in Down syndrome pregnancythan in normal pregnancy15, the final degra-dation product of hCG in urine, bCF, mightbe even more increased in Down syndromepregnancy than hCG itself. Although thereare no data, as yet, to support a specificmechanism for the increased nicking of hCGin Down syndrome pregnancy, we canspeculate that a trisomic placenta is a moreforeign allograft than a normal placenta andtherefore it might evoke a greater phagocyticattack with increased proteolytic activity, in-cluding specific hCG-nicking activity.

Many of the early studies that were pub-lished strongly supported the hypothesis ofCuckle et al., with findings that showed thaturine bCF levels are 5–6-fold higher inDown syndrome pregnancies than in unaf-fected ones16,17; for comparison, maternalserum hCG and the free b subunit of hCG arein the range of 2.0–2.3 fold higher in Downsyndrome pregnancies. However, the averageincrease does not tell the full story. The distri-bution of bCF values in affected versus unaf-fected pregnancies was such that 65–88% ofthe bCF values in Down syndrome pregnan-cies were above the 95th centile of unaffected values, making bCFthe single, best screening marker that has been studied to date.

However, as more studies have appeared in the literature, the earlyresults have proven to be on the high end of the spectrum. In fact, the

variability of results thus far obtained has been unusually great (Fig.3). Presently, the range of the median increase of bCF concen-trations in Down syndrome pregnancy is between 1.3 and 6.1 timesthe unaffected median, with an average of 4.1 times the median7.Such variability indicates that the average might not be the final cor-rect estimate. Rather, it indicates that other factors, which are notreadily identifiable, might be affecting the results of studies. It couldbe that the final answer might be at one extreme of this range or theother, rather than in between.

To add to the uncertainty surrounding the potential of bCF as ascreening marker, the results of two new studies will soon be avail-able. One, an international collaborative trial, was organized in 1996to resolve data discrepancies with respect to the use of bCF as amarker in prenatal screening for Down syndrome (H.S. Cuckle et al.,unpublished). In total, 6731 women from eight centers were tested be-tween 14 and 19 completed weeks of gestation. Of these, 4219 werelow-risk women who were tested at the time of routine maternalserum screening, and 2512 were high-risk women who were tested atthe time of amniocentesis for reasons other than positive serumscreening results. There were seven cases of Down syndrome found inthe low-risk group and 32 were identified in the high-risk group. Theresults were disappointing. The median value in the 39 cases of Downsyndrome was only 1.7 times the unaffected median (95% confidenceinterval, 1.2–2.2). Nine of the 39 cases (23%) exceeded the 95th cen-tile among unaffected pregnancies. The results were similarly poor forboth the low- and the high-risk components of the trial.

The other study18 is a continuation of a study that was originallypublished by Isozaki et al.19 and that was carried out at Yale

71

R e v i e w sMOLECULAR MEDICINE TODAY, FEBRUARY 1999

Figure 2. The pathway of estriol biosynthesis and excretion during pregnancy. The green pathway indi-cates the primary estriol biosynthetic route, which uses fetal precursors almost exclusively. The blue path-way shows the biosynthesis of other estrogens, which uses both maternal and fetal precursors, duringpregnancy. Unconjugated estriol, estrone and estradiol are conjugated to estriol sulfates and glucuronidesin the maternal liver. The chemical structure shows unconjugated estriol [3,16a,17b-trihydroxyestra-1,3,5(10)-triene]. DHEAS, dehydroepiandrosterone sulfate.

Fetus

Cholesterol

DHEAS

16α-hydroxyDHEAS

Metabolized inadrenal gland

Metabolizedin liver

DHEAS

Estrone and estradiol

16α-hydroxyDHEAS

Estriol

Cholesterol

DHEAS

Estrone and estradiol

Unconjugated estriol

Estriol sulfatesand glucuronides

Urinary excretion

Placenta Mother

Metabolized inmaternal liverOH

OH

HO

Glossaryb-core fragment (bCF) – The major metabolite of human chori-onic gonadotropin (hCG) that is found in maternal urine.

Genetic amniocentesis – The procedure by which the fluid andcells that surround a developing fetus are withdrawn for genetic orcytogenetic analysis.

bCF monoclonal sandwich assay – A two-step assay in which amonoclonal capture antibody against b-core fragment (bCF) is an-chored to a microtitre well or to a polystyrene tube. Following incu-bation with urine and washing, a second signal antibody againstthe b subunit of human chorionic gonadotropin (hCG), which isconjugated to horse radish peroxidase, is applied to the well andproduces a colored substrate.

Nondysjunction – The failure of homologous chromosomes orchromatids to separate during meiosis or mitosis.

Second trimester – Weeks 14–26 of pregnancy; prenatal screen-ing is performed during the first part of this time span.

University. This study has continued for three years, and 1157women have contributed urine samples at the time of amniocentesis(that was performed on the basis of their age); 23 cases of Down syn-drome were identified. The median urinary bCF level in affectedpregnancies was, on average, 5.4 times the unaffected level, with65% above the 95th centile of unaffected. Results were consistentover the three-year study period, with 63%, 66% and 66% of casesexceeding the 95th centile in each of the three years. Thus, in con-trast to the multicenter trial, this study has had, and continues tohave, excellent results with bCF.

Although different bCF assays were usedin each of the two latest studies, this is un-likely to have caused the difference in re-sults, because it has been previously shownthat the two assays are almost identical intheir specificity20. Both are monoclonalsandwich-type assays that have the sameanti-bCF capture antibody and differentanti-b subunit signal antibodies. One differ-ence that requires further examination is thatthe multicenter trial froze, short-term, manyof its samples, whereas the Yale study18 meas-ured samples that were never frozen, on aweekly basis. Studies by Birken et al. havedemonstrated that bCF is present in urine as monomers and aggregates21. It has been recently shown by Cole et al.18 thaturine samples that contain high titres of bCF tend to lose immunoreactivity uponstorage, supporting the concept that bCF hasa greater tendency to aggregate at higherconcentrations.

The hyperglycosylated variant of hCG hasbeen recently examined22,23. Using two dif-ferent, specific monoclonal immunoassays,Cole et al. tested urine samples from .1000controls and 23 Down syndrome pregnan-

cies, and found nine of ten cases in the first study22 and 18 of 23cases (79%) in the second study23 above the 95th centile of controllevels. These results indicate that particular molecular forms of hCGmight be expressed preferentially in Down syndrome pregnancy.Confirmatory studies are now in progress in other centers to deter-mine whether these encouraging results are correct.

In addition to bCF, the maternal urine levels of tE3 have been ex-amined in Down syndrome pregnancy. Levels of urinary tE3 are lowin affected pregnancies, as are the levels of uE3 in maternalserum17,24,25. In three studies published to date, the median tE3 levelin Down syndrome pregnancy is about 0.65 times that of the unaf-fected median, with between 22% and 35% of cases below the 5thcentile of unaffected pregnancy levels. If maternal urine screeningwith bCF becomes a reality, then it is likely that urinary tE3 will bemeasured as well.

Dilemmas in the bCF findingsThe extreme variability of the studies on urinary bCF levels inDown syndrome pregnancies suggests that factor(s) that are not yetfully recognized either exaggerate or mask the differences in levelsbetween affected and unaffected pregnancies. A number of possi-bilities require further examination. For example, all but one of thepublished studies on urinary bCF were based on the retrospectiveanalysis of frozen samples, and all the studies used Down syndromecases that were mainly derived from high-risk pregnancies. Somewomen might have been tested because they had abnormal maternalserum screening results; this would bias urinary bCF findings be-cause of the correlation between levels of maternal serum hCG andfree-b hCG, and maternal urine bCF. However, the two, as yet, un-published studies were prospective; one, in part, examined low-riskpregnancies and both excluded women who were considered to be ata high risk on the basis of maternal serum screening.

72

R e v i e w s MOLECULAR MEDICINE TODAY, FEBRUARY 1999

Figure 3. Median second trimester urinary b-core fragment levels in Down syndrome pregnancies fromseven published and two unpublished studies. Asterisks denote the two unpublished studies. The verticalbars are the 95% confidence intervals. A horizontal line is drawn at 1.0 MOM, the unaffected median. Adashed line depicts the consensus median for Down syndrome cases (3.35 MOM). Study 1 (Ref. 17) n 523; study 2 (Ref. 18) n 5 23; study 3 (Ref. 16) n 5 14; study 4 (Ref. 26) n 5 18; study 5 (Ref. 19) n 5 12;study 6 (Ref. 27) n 5 29; study 7 (Ref. 28) n 5 28; study 8 (H.S. Cuckle et al., unpublished) n 5 40; study9 (Ref. 29) n 5 5. MOM, multiple of the median.

10.0

1.0

0.1

Urin

ary

β-co

re fr

agm

ent (

MO

M)

1 2 3 4 5 6 7 8 9

Study

*

*

The outstanding questions

• What is the biological basis for the increase in b-core frag-ment (bCF) and other human chorionic gonadotropin (hCG)metabolites in Down syndrome pregnancies?• Is the level of bCF raised in women who carry other tri-somic fetuses, and could it be used to screen for these too?• Are some forms of bCF, or other hCG metabolites, el-evated more than others in the urine of women who carryDown syndrome fetuses, and could this form the basis of amore accurate assay?• What factor or factors have caused the enormous variationin urine levels of bCF that have been reported in differentclinical trials? Will it be possible to eliminate all these factorsto develop an acceptable alternative to serum screening?• Is there a combination of urine and serum markers that willprovide a high detection rate while substantially reducing thefalse positive rate in screening for Down syndrome?

An additional problem could be that most of the studies used ahighly sensitive assay that was originally designed to detect bCFlevels in the fmolml21 concentration range. In contrast, urine frompregnant women has bCF levels that are, on average, 10 000 foldhigher than this range, requiring the extensive dilution of each urinesample prior to testing. This introduces an additional source of errorinto bCF distribution values in a population of pregnant women, al-though it would not be expected to bias the results in a particular di-rection. Two other possible sources of variability might, in fact, biasresults. Urinary analyte results are generally normalized by measur-ing the creatinine level, but this might not completely normalize uri-nary analyte levels that are found in extremely concentrated or dilutespot urine samples. Additionally, bCF levels are markedly affectedby the time of day that urine is collected – levels are highest in themorning and lowest in the evening. Such sources of variability havethe potential to bias results, although they are probably not the ulti-mate reasons for the variability in the results. Finally, it is difficult, ifnot impossible, to assess the effect of publication bias against studiesthat have negative or poor results, both by authors who are reluctantto submit these results for publication and by referees who might bereluctant to accept them.

Concluding remarksPrenatal screening for Down syndrome in the second trimester ofpregnancy using multiple serum markers has been available for al-most ten years. Currently, four serum markers are used to achieve adetection rate of ~80%, through the identification of 5% of the preg-nant population as being at high risk, although most clinical labora-tories are not in a position to measure four markers. Early studies onscreening using maternal urine were very encouraging and promisedimproved detection rates and a less expensive and simpler test.However, the results of subsequent studies have shown great vari-ability. Presently, it is not at all certain whether maternal urinescreening will, in fact, become a clinical reality. It still has potential,but the true clinical value of maternal urine screening remains to bedetermined.

References1 Tabor, A. et al. (1986) Randomised controlled trial of genetic amniocentesis in

4606 low-risk women, Lancet i, 1287–12932 Merkatz, I.R., Nitowsky, H.M., Macri, J.N. and Johnson, W.E. (1984) An associ-

ation between low maternal serum alpha-fetoprotein and fetal chromosomalabnormalities, Am. J. Obstet. Gynecol. 148, 886–894

3 Cuckle, H.S., Wald, N.J. and Lindenbaum, R.H. (1984) Maternal serum alpha-fetoprotein measurement: a screening test for Down’s syndrome, Lancet i,926–929

4 Bogart, M.H., Pandian, M.R. and Jones, O.W. (1987) Abnormal maternalserum chorionic gonadotropin levels in pregnancies with fetal chromosomeabnormalities, Prenatal Diagn. 7, 623–630

5 Macri, J.N. et al. (1990) Maternal serum Down syndrome screening: free b-protein is a more effective marker than human chorionic gonadotropin, Am.J. Obstet. Gynecol. 163, 1248–1253

6 Canick, J.A. et al. (1994) Maternal serum dimeric inhibin is elevated in Downsyndrome pregnancy, Am. J. Hum. Genet. 55, A9

7 Canick, J.A. et al. (1988) Low second trimester maternal serum unconjugatedoestriol in pregnancies with Down’s syndrome, Br. J. Obstet. Gynaecol. 95,330–333

8 Wald, N.J., Kennard, A., Hackshaw, A. and McGuire, A. (1997) Antenatalscreening for Down’s syndrome, J. Med. Screening 4, 181–246

9 Wald, N. et al. (1996) Prenatal screening for Down’s syndrome using inhibin-A as a serum marker, Prenatal Diagn. 16, 143–153

10 Yen, S.S.C. (1994) Endocrinology of pregnancy, in Maternal Fetal Medicine,Principles and Practice (3rd edn), (Creasy, R.K. and Resnik, R., eds),pp. 382–412, Saunders

11 Cole, L.A. et al. (1991) The heterogeneity of hCG. III. The occurrence, bio-logical and immunological activities of nicked hCG, Endocrinology 129,1559–1568

12 Cole, L.A., Kardana, A., Park, S.Y. and Braunstein, G.D. (1993) The deacti-vation of hCG by nicking and dissociation, J. Clin. Endocrinol. Metab. 76,704–710

13 Elliott, M.M., Kardana, A., Lustbader, J.W. and Cole, L.A. (1997) Carbohydrateand peptide structure of the a- and b-subunits of human chorionic go-nadotropin from normal and aberrant pregnancy and choriocarcinoma,Endocrine 7, 15–32

14 Cuckle, H.S., Iles, R.K. and Chard, T. (1994) Urinary b-core human chorionicgonadotropin: a new approach to Down’s syndrome screening, PrenatalDiagn. 14, 953–958

15 Rotmensch, S. et al. (1996) Peptide heterogeneity of human chorionic go-nadotropin (hCG) and its b-subunit in Down syndrome pregnancies, Am. J.Obstet. Gynecol. 166, 354–360

16 Canick, J.A. et al. (1995) Second trimester levels of maternal urinary go-nadotroin peptide in Down syndrome pregnancy, Prenatal Diagn. 15, 752–759

17 Cuckle, H.S. et al. (1995) Urinary multiple marker screening for Down’s syn-drome, Prenatal Diagn. 15, 745–751

18 Cole, L.A. et al. New screening tests for fetal Down syndrome: I. Fresh urineb-core fragment, Prenatal Diagn. (in press)

19 Isozaki, T., Palomaki, G.E., Bahado-Singh, R.O. and Cole, L.A. (1997)Screening for Down syndrome using b-core fragment: prospective study,Prenatal Diagn. 17, 407–413

20 Cole, L.A. et al. (1997) Comparison of 12 assays for detecting hCG and re-lated molecules in urine samples from Down syndrome pregnancies, Prenat.Diagn. 17, 607–614

21 Birken, S. et al. (1988) Structure of human chorionic gonadotropin b-subunitcore fragment from pregnancy urine, Endocrinology 123, 572–580

22 Cole, L.A. et al. (1998) Hyperglycosylated hCG, a potential alternative tohCG in Down syndrome screening, Prenatal Diagn. 18, 926–933

23 Cole, L.A. et al. New screening tests for fetal Down syndrome: II.Hyperglycosylated hCG, Prenatal Diagn. (in press)

24 Kellner, L.H. et al. (1997) Levels of urinary beta-core fragment, total oestriol,and the ratio of the two in second-trimester screening for Down syndrome,Prenatal Diagn. 17, 1135–1142

25 Cole, L.A. et al. (1997) Combining b-core fragment and total oestriol meas-urement to test for Down syndrome pregnancies, Prenatal Diagn. 17,1125–1134

26 Canick, J.A. et al. (1996) A second level evaluation of maternal urinary go-nadotropin peptide as a marker for second trimester Down syndrome screen-ing, in 3rd Joint Clinical Genetics Meeting, A156, March of Dimes Birth DefectsFoundation and the American College of Medical Genetics

27 Lam, Y.H. et al. (1997) Second-trimester maternal urinary gonadotrophinpeptide screening for fetal Down syndrome in Asian women, Prenatal Diagn.17, 1101–1106

28 Spencer, K., Aitken, D.A., Macri, J.N. and Buchanan, P.D. (1996) Urine freebeta hCG and beta core in pregnancies affected by Down’s syndrome,Prenatal Diagn. 16, 605–613

29 Hayashi, M. and Kozu, H. (1995) Maternal urinary beta-core fragment ofhCG/creatinine ratios and fetal chromosomal abnormalities in the secondtrimester of pregnancy, Prenatal Diagn. 15, 11–16

73

R e v i e w sMOLECULAR MEDICINE TODAY, FEBRUARY 1999