Chromosome Abnormalities and Genetic Counseling , 3rd Edition

23.Prenatal Diagnostic Procedures

The means to diagnose the fetal karyotype has provided medical cytogenetics with one of its major areas of application. The discovery of an abnormality allows the option of termination of the pregnancy or, later in gestation, a more suitable obstetric management. The main indications for prenatal cytogenetic diagnosis are the following: (1) the pregnant woman being of advanced childbearing age, (2) parental heterozygosity for a chromosome rearrangement, (3) the birth of a previous child with a chromosome defect, (4) abnormal maternal blood biochemistry, and (5) fetal anomaly detected on ultrasonography.

Fetal Ultrasonographic Anomalies

The issues of maternal age, parental chromosome rearrangement, previous history, and maternal serum screening are dealt with in other chapters; a few comments on ultrasonographic screening are appropriate here. Ultrasonography now has a considerable role in prenatal diagnosis. In Victoria, Australia, for example, 20% of prenatal chromosome tests in 1999 were done on the grounds of ultrasound findings of a fetal malformation or of a marker of aneuploidy (Webley and Halliday, 2000). Certain major ultrasonographic defects are fairly specific: for example, holoprosencephaly predicts the likelihood of trisomy 13, fetal hy-drops/cystic hygroma predicts monosomy X or trisomy 21, and an endocardial cushion defect or duodenal atresia predicts trisomy 21. The minor marker of increased nuchal translucency (actually, this separation of the skin from the underlying tissue can extend from as far as the occiput down to the lower back) is less specific. Cardiac malformations generally have a frequent association with fetal aneuploidy, as do certain renal defects (Yates, 1999; Amor et al., 2003). Daniel et al. (2003a) reviewed 1800 cases in which an anomaly (an actual malformation, or a minor marker of aneuploidy) had been detected at ultrasonography, and assembled a table of risks of aneuploidy according to the findings (Table 23-1). The abnormal karyotypes included trisomies 13, 18, and 21, triploidy, 45,X and mosaics, various autosomal and gonosomal duplications and deletions, rare trisomies, and de novo apparently balanced rearrangements. The sign of increased nuchal translucency indicates an increased risk for fetal trisomy, the level of risk proportional to the degree of separation, and this observation is sufficiently robust that is used for pregnancy screening of aneuploidy (Chapter 22). Cysts of the choroid plexus (tissue within the cerebral ventricles) are a soft marker for trisomy 18, but not trisomy 21 (Walkinshaw, 2000). On the specific question of rare autosomal abnormalities (rare trisomies, deletions, duplications, supernumerary markers, various other structural rearrangements), a large European series based upon reports from malformation registers in several jurisdictions linked ultrasound findings to cytogenetic results (Baena et al., 2003). Nearly half of all rare autosomal abnormalities showed fetal anomalies on ultrasonography, with heart and brain defects and growth retardation more often seen with deletions, and cystic hygroma, hydrops and nuchal translucency more typically associated with trisomies and duplications. These rare abnormalities comprised 7% of all chromosomally abnormal prenatal diagnoses.

Table 23.1. Likelihood of Discovering a Chromosome Abnormality at Prenatal Diagnosis, According to Pattern of Defects Identified at Fetal Ultrasonography, for All Maternal Ages

Defects

Likelihood of aneuploidy
(%)

CNS/cranial shape plus cardiaca

51

Key malformation,b singly or in combination

37

CNS ± otherc

20

Nuchal translucency, first trimester, ± other abnormality

25

Nuchal thickening, second trimester, ± other abnormality

14

Cardiac ± other abnormality

11

Pyelectasis/two vessel cord/echogenic bowel/short femur

6

Other (singly or in combination)

3

Some percentages are considerably higher or lower for older and younger maternal ages, respectively. These data were obtained prospectively. From Daniel et al. (2003a).
CNS, central nervous system; IUGR, intrauterine growth retardation.
aExcluding anencephaly/spina bifida, including choroid plexus cysts.
bCystic hygroma/hydrops/exomphalos/severe IUGR/duodenal atresia/talipes.
cExcluding anencephaly/spina bifida/cardiac.

There are some subtleties in the choice of language in this setting, as de Crespigny et al. (1996, 1999) discuss. We speak of the pregnant woman as a mother, yet she is not; neither is her husband/partner as yet a father. Equally, the fetus is not a baby, not acquiring that status until ex utero existence is achieved. But of course many parents-to-be, not to mention professionals (including us), use these words. Counselors should be sensitive to these subtleties. De Crespigny observes that, if an ultrasonologist should discover a fetal defect, using the terms “baby” and “mother” may exert indirect pressure on the couple to continue the pregnancy. “Although many women regard a fetus as a baby from the very beginning, others will be affronted if their doctor does not seem to recognize this difference between a fetus and a baby, which they may interpret as interfering with the pregnant woman's reproductive freedom.” As always, counselors will need to know their patients, and judge the right words to use and the way to say them (Benkendorf et al., 2001).

PRENATAL DIAGNOSTIC PROCEDURES

Since the early 1970s, prenatal diagnosis (PND) of chromosome disorders has been done by culture of amniotic fluid cells obtained by amniocentesis at about 16 weeks of pregnancy. A number of other approaches to PND have since been developed, ranging from preimplantation diagnosis (following in vitro fertilization), through chorion villus sampling (CVS), to fetal blood sampling, and some more experimental procedures. Naturally, parents-to-be are anxious to have results as early as possible. A desire for an early result needs to be balanced against a number of considerations which can include complexity of the procedure, both clinically and in the laboratory, procedural trauma and risks, reliability of results, cost, and the prior risk for a fetal abnormality. A useful source for the lay person is Prenatal Testing. Making choices in Pregnancy (De Crespigny et al., 1998).

Quality of Cytogenetic Results. Blood lymphocyte culture followed by high-resolution banding gives the best-quality cytogenetics. The quality from any other tissue is almost never as good. Mostly, this is not a serious problem in prenatal diagnosis, since this is primarily aimed at detecting complete aneuploidy. It is important that both patient and doctor realize that very subtle chromosome abnormalities, which could have a serious consequence, may escape detection, regardless of the procedure chosen. In the absence of a known risk factor, microdeletion syndromes are unlikely to be detected at routine prenatal diagnosis. Of course, if a subtle parental rearrangement is known, the cytogeneticist knows precisely what to look for and can apply a specific focus, possibly resorting to molecular cytogenetics to provide a clear picture.

Blastomere and Polar Body Biopsy

The techniques of preimplantation genetic diagnosis (PGD) have advanced very considerably in recent years, and a separate chapter (Chapter 24) is now devoted to a treatment of this category of prenatal diagnosis, in which “prenatal” can be taken to mean 37½/ weeks prior to birth.1

Chorionic Villus Sampling

Chorionic villus sampling is typically a first-trimester procedure, the usual time being at 10–11 weeks gestation. (The expression “placental biopsy” could also be applied, although in practice this term is used when the testing is done in later pregnancy; see below). The earlier period of diagnosis permitted by CVS may be seen as more useful in the setting of a higher genetic risk. If a genetic abnormality is identified and abortion is chosen, this can be, prior to 14 weeks, a more private matter, and the termination procedure is an operative intervention (curettage or suction evacuation of the uterus). Couples are more likely to make a choice for abortion when the diagnosis has been made in the first trimester (Verp et al., 1988). There is potential in CVS for diagnostic difficulty due to the occasional detection of confined placental mosaicism (Chapter 25). Nonmosaic results for the common aneuploidies are, however, highly reliable (Smith et al., 1999b). In experienced hands there is a high degree of safety, and the risk of procedure-related miscarriage is usually given as 1.5–2% (Brambati et al., 2002; Brun et al., 2003). In one study in which lethal fetal diagnoses were accounted for, the loss rate up to 2 weeks post-procedure was only 0.23% (1/436) (Nanal et al., 2003).

Direct, Short-term, and Long-term Chorionic Villus Sampling. Chorionic villi can be analyzed directly (same day), after short-term culture (next day or two), or after long-term (a week or so) culture. Trophoblast is the source of the cell population studied at direct and short-term CVS culture. These cells are no longer extant (if they have not already been removed by trypsinization at sample receipt) after the first few days, and it is the mesenchymal core of the villus that provides the cells that are analyzed at long-term culture (and see Color Fig. 25-1).

In the early 1990s there were disconcerting reports of an increased incidence of transverse limb deficiencies and tongue and jaw defects—“oromandibular-limb hypogene-sis”—following early CVS (before 10 weeks, and especially up to 8 weeks). The association appears likely to be causal, and one line of circumstantial evidence is that the rate of anomalies falls with increasing gestational stage from 9 to 11 weeks (Firth, 1997). Various mechanisms have been proposed: oligohydramnios, bradycardia, hypovolemia, thromboembolism, vasoconstriction, antibody-me-diated reaction, and increased apoptotis following disruption of end arteries (Luijsterburg et al., 1997). Given these observations, there is a consensus that CVS should not be done earlier than 10 weeks.

Standard Amniocentesis

Transabdominal amniocentesis, done at about 15–17 weeks gestation, has been the standard cytogenetic prenatal diagnostic procedure for over a quarter of a century. It has a high degree of safety to both mother and fetus; maternal complications, or fetal injury due to direct trauma, are practically unknown. The risk for maternal Rhesus immunization (Rh-negative mother, Rh-positive fetus) can be circumvented by administering an antibody injection. The only significant complication is a procedure-related fetal loss rate of about 0.5%. The cytogenetic results are highly reliable. The biological sources of error are, first, that maternal rather than fetal cells, or a mixture of both, are sampled. In practical terms, this rarely causes a problem. Second, fetal mosaicism may go undetected, since only a limited number of cells can feasibly be examined. Very few examples of this error are recorded.

Amniotic fluid culture has a high success rate. Persutte and Lenke (1995) have suggested that if amniotic cells fail to grow, for no obvious reason, there may be a substantial risk for fetal aneuploidy (13% of 32 cases in their preliminary study). This assessment was supported in a large systematic study from London (Reid et al., 1996), in which 42 failures (1%) among 4134 amniocenteses were followed up. Complete information was obtained on all but 1 of these 42 cases. Karyotyping was ultimately done in most (78%) of these failed cases and of these, 19% revealed an abnormality (comparing with a 4% abnormality rate in the whole material). The clear lesson from these studies is that women having had a failed amniocentesis culture should be offered careful review and retesting.

The obvious disadvantage of standard amniocentesis is that the results are not available until about 16–18 weeks. If the reason for the amniocentesis was an increased-risk result from maternal serum screening, the procedure may not be done until 17–18 weeks, aggravating this difficulty. If a result cannot be available by 20 weeks, another procedure (placental biopsy, interphase FISH) may be worth considering.

Early Amniocentesis

In the late 1980s, early (10–13 weeks) amniocentesis was proposed as an alternative to CVS. In a carefully controlled comparison, Nicolaides et al. (1994)found a 2%–3% additional fetal loss rate in early amniocentesis and, possibly,  a higher incidence of talipes amongst subsequently born children. Daniel et al. (1998) compared 10- to 14-week procedures with 15 weeks and upward, and observed that the early amniocentesis samples were not quite as satisfactory, multiple needle insertions were more often required, and the pregnancy loss rate was greater. On the whole, the differences were not great, other than the loss rate of 2.2% in the early group compared with only 0.6% in the midtrimester group. Similar figures are reported by Collins et al. (1998). In the Canadian Early and Midtrimester Amniocentesis Trial, the findings for 11 weeks through to 13 weeks, 6 days were somewhat more disconcerting, with more complications and a higher culture failure rate (Delisle and Wilson, 1999). Indeed, Jauniaux et al. (2000) have called for the procedure to be abandoned because of the risk of fetal structural defects.

Fetal Blood Sampling

Fetal blood is aspirated by direct puncture of a blood vessel, usually in the umbilical cord (cordocentesis), and a cytogenetic result is obtainable after a few days. Before FISH analysis of uncultured cells (see below) came to be more widely used, cordocentesis was useful when speed of diagnosis was of the essence, in the setting of the detection of a fetal anomaly at ~18 week ultrasonography. The procedure once had a role in assisting resolution of mosaicism in amniotic fluid culture (Shalev et al., 1994) but this has largely been replaced by the use of FISH.

Placental Biopsy

In principle, this is the same as first-trimester CVS. The placenta is sampled by a transabdominal approach, and this is a straightforward procedure. The main application had been when a rapid result was needed, although FISH methodology (see below) can now largely accommodate that imperative. An insufficient amount of amniotic fluid remains an indication.

Experimental Approaches

A variety of experimental approaches to prenatal diagnosis have been, or continue to be, under trial, and for the sake of completeness we mention them here. None is likely to replace amniocentesis or CVS, at least in the near future.

Fetal Cell and DNA Isolation from Maternal Blood

The two important cell types that are released from fetal tissue into the maternal circulation are the nucleated red blood cell and the trophoblast. Various means, none of which is very satisfactory, are employed to separate out these components from a sample of maternal blood (Wachtel et al, 2001). Multicolor FISH or molecular methodologies may be applied to the cells for detection of the major aneuploidies.

The Nucleated Red Cell. Sophisticated methodologies are necessary in order to separate the fetal from the maternal nucleated red blood cells (NRBCs). A major collaborative research study is ongoing, under the aegis of the U.S. National Institutes of Health, comparing NRBC results with those of amniocentesis or CVS (Bianchi et al., 2002). According to the data from the first 5 years, false-positive rates for aneuploidy on FISH analysis are low, mostly about 1%. But the sensitivity is low and, according to the criteria of analysis, around a quarter of trisomies may be missed.

While the applicability of maternal blood sampling is for population screening, and at an early gestational stage, Wang et al. (2000) have described its use in a different setting. A woman heterozygous for a translocation t(1;6)(p31;q14) had previously had six miscarriages, and was unwilling to take the risk of an invasive prenatal diagnostic procedure. A blood sample was drawn at 20 weeks gestation, and presumed fetal cells extracted. Analysis by FISH with no. 1 and no. 6 probes indicated a disomic fetus. A 46,XX child was born. The reader will note that this is a translocation characterized by long centric and translocated segments (Fig. 4-5d), and so any unbalanced combination would inevitably have miscarried. That being so, an apparently normal fetus on ultrasonography at mid-pregnancy could rather confidently have been predicted to have a normal or balanced karyotype. Thus, the particular value of this study was to show the feasibility, potentially, of this procedure.

The Trophoblast Cell. Trophoblast cells invade the decidua and myometrium, migrating into the uterine blood vessels, during early placentation (Oudejans et al., 2003). Some enter the maternal circulation, and the concentration in maternal blood is about 1 cell per milliliter. Enrichment procedures (which obviously must be very efficient) allow separation of the trophoblast cells. Another approach is to use as the source material fetal DNA that is present in maternal plasma, taking advantage of the fact that trophoblast cells that have undergone apoptosis release their DNA into the maternal circulation.

Cervical Lavage

Trophoblast cells may migrate from the confines of the uterine cavity and enter the endocervical canal, and can be collected for molecular analysis by endocervical irrigation and aspiration (lavage) (Bussani et al., 2002). The point is to be noted that lavage may not necessarily bypass the risk for limb defects associated with early CVS, although the detail of how the procedure is performed may be the key element (Chou et al., 1997; Daryani et al., 1997; Hsi and Adinolfi, 1997).

Cervical Cytobrush

Papanicolaou cervical sampling (“Pap smear”) is a routine procedure that has been used for decades. Might it be successful as a prenatal diagnostic technique, exploiting the fact that some trophoblast cells migrate down as far as the region of the external os? These would presumably be the same population of cells as those which are tested at direct CVS culture. Potentially, diagnosis as early as 7 weeks would be feasible. Fejgin et al. (2001) have tested the concept using FISH, with tentatively promising results, but in the trial reported in Cioni et al. (2003) neither FISH nor a molecular approach was satisfactory. Single-cell molecular methodology (Findlay et al., 2001) may be more accurate, and I. Findlay (pers. comm., 2003) is assessing this technique, using a panel of markers interrogating the more important aneuploidies, having applied an enrichment procedure to harvest the few fetal cells from the sample. In principle, trisomy would be detectable on the observation of three alleles at a marker locus.

Cystic Hygroma and Pleural Effusion Fluid

Cystic hygroma has a strong association with fetal aneuploidy, especially monosomy X. A concomitant oligohydramnios may make amniocentesis difficult. Fluid from cystic hygroma and pleural effusion contains lymphocytes and these cells can be cytogenetically analyzed within the time frame of a few days. In one small series, three out of four cystic hygroma analyses showed aneuploidy (trisomy 21, monosomy X) (Costa et al., 1995).

Celocentesis

The extraembryonic celom (Color Fig. 25-1) is a source for cells originating from extraembryonic mesoderm, to which a FISH analysis can be applied (Crüger et al., 1997a). The procedure has the attraction of an earlier timing (7–9 weeks) than CVS. One report exists of prenatal diagnosis of monosomy X by this means (Crüger et al., 1997b). But a high post-procedure miscarriage rate is likely to stifle much further interest (Ross et al., 1997).

Newer Analytical Methodologies Applied to Amniocentesis and Chorionic Villus Sampling

Fluorescence in Situ Hybridization

This method can be applied to interphase nuclei with chromosome-specific probes to detect a single, double, or triple dose of a particular chromosome (Morris et al., 1999). Thus the need for culture is bypassed, regardless of whether the cells were from amniotic fluid, CVS, or fetal blood, offering earlier diagnosis. A FISH-based result can be given within the space of 1 working day. This approach is particularly useful where the need for a rapid result is more pressing, such as when fetal anomalies have been discovered at ultrasound. In one small series with particular reference to the third trimester, Aviram-Goldring et al. (1999)showed an aneuploidy in 23% of pregnancies with intrauterine growth retardation and structural abnormalities: five with trisomy 21, and two with trisomy 18. Feldman et al. (2000) similarly applied amniotic fluid cell FISH to high-risk pregnancies (that is, with ultrasonographic abnormalities). They detected 14 cases of trisomy 21, 10 of trisomy 18, 3 of trisomy 13, 4 of monosomy X, and 1 of triploidy in 4193 samples over the period 1996–1998, for a total abnormality rate of 10.6%. A limitation is that only those specific aneuploidies being tested for (the major autosomal trisomies and sex chromosome aneuploidies) could be detected, and it is fair to say that conventional karyotyping is not about to be replaced by FISH (Morris et al., 1999). Waters and Waters (1999) comment that, with shortening times for standard karyotyping, there may actually be less pressure for a rapid result; nevertheless, the attraction of being able to offer a next-day result, albeit a preliminary one, is clear.

The question of false-negative results arises. Weremowicz et al. (2001) reviewed their experience over 1992–2000, during which time they applied FISH to some 8% of the 11,000 amniocentesis samples coming into their laboratory for routine karyotyping, this 8% including cases with an increased risk (abnormalities on ultrasound, serum screen result). In the whole material, there were 89 abnormalities identified on the routine karyotypes that would have been expected to be picked up by FISH (using probes for 13, 18, 21, X, and Y). In the event, 75 (84%) of these were found. The missed cases included eight with an inconclusive result, one with no result, and, more importantly, five false-negatives. Of these latter five, the true karyotypes were trisomy 18 (two cases) and trisomy 21 (three cases). Technical problems related to poor hybridization efficiency (low copy number of the DNA repeats being probed, for example) and maternal blood contamination of the fluid sample are plausible explanations. Like others, Weremowicz et al. note the usefulness of the FISH approach in being able to provide a rapid answer particularly when there are grounds for suspecting an abnormality or if the pregnancy is more advanced, but they also emphasize the need for careful counseling so that patients are aware of the limitations. With respect to trisomy 21, Witters et al. (2002) had a encouraging record: in a similar study comprising 5049 amniotic fluid samples, in which interphase FISH was applied in parallel with conventional karyotyping, all 70 cases of trisomy 21 were detected, and no false-positive result arose. One false-positive is on record, however, probably due to technical aspects of probe hybridization (George et al., 2003). On the question of mosaicism, Van Opstal et al. (2001) note that FISH on uncultured cells may provide a more accurate picture than with cultured cells, the latter possibly being subject to selective pressure in vitro and the abnormal cells more prone to fail. On the other hand, the class of amniocyte that grows preferentially in culture (namely, amniotic mesoderm) might, according to the reinterpretation of Robinson et al. (2002), more closely reflect the true embryonic state.

Focused FISH can be applied in specific circumstances. The ultrasound discovery of a cardiac outflow tract abnormality would, for example, point to the need for 22q11 analysis. A rapid diagnosis is particularly to be desired in the setting of parental heterozygosity for a chromosome rearrangement, in which there may be a high risk for abnormality, and FISH can provide this. Thus, Cotter and Musci (2001) used subtelomeric probes for 5pter, 5qter, and 14qter to enable rapid diagnosis for a pregnant woman with the karyotype 46,XX,t(5;14)(p14.2;p13) who had had a previous child with cri du chat syndrome. Similarly, Pettenati et al. (2002) applied this approach in the setting of parental heterozygosity for a number of reciprocal or Robertsonian translocations.

Molecular Methodology

DNA-based tests may enable rapid diagnosis. Pertl et al. (1999) and Solassol et al. (2003) have undertaken preliminary studies using the approach of quantitative fluorescence PCR, with chorionic villus samples and uncultured amniocytes respectively being the test material. The sequences assessed are chromosome-spe-cific polymorphic small tandem repeats. The fact that there are three allelic forms of a particular marker allows the assumption of trisomy for that chromosome. The use of comparative genomic hybridization (CGH) to assess arrays of sequences, specific either for known microdeletion syndromes or subtelomeric regions, or across the whole genome, is being developed (see p. 12).

Primum non nocere

“First, do no harm” is a cornerstone of medical practice. Yet, almost inevitably, having a prenatal diagnostic procedure causes anxiety. Rothman (1988), in her book The Tentative Pregnancy, is particularly critical of what she sees as a medicalized distortion of the normal process of being pregnant. Hodge (1989) describes her personal experience in Waiting for the Amniocentesis, and we reproduce her letter in full:

I drafted the following letter to the editor one week before I expected to hear the results of my amniocentesis:

“I am 40 years old and 19 weeks pregnant with what will presumably be my third child. I am on the basic science faculty of a medical school. When I teach medical students about amniocentesis, I occasionally mention the difficulty for the woman of having to wait until well into the second trimester to receive her results.

“I am in that situation myself now, awaiting my results. And before experiencing it, I was unprepared for two phenomena. One was just how difficult the wait is. Pregnancy is always a time of waiting, but now time has slowed down to an extent I did not anticipate. The other, more disturbing phenomenon is how the waiting has affected my attitude toward the pregnancy. At many levels I deny that I really am pregnant “until after we get the results.” I ignore the flutterings and kicks I feel; I talk of “if” rather than “when” the baby comes; I am reluctant to admit to others that I am pregnant. I dream frequently and grimly about second-trimester abortions. In some sense I am holding back on “bonding” with this child-to-be. This represents an unanticipated negative side effect of diagnostic amniocentesis. And all this, even though my risk of carrying a chromosomal abnormality is less than 2 percent.

“I presume I am not alone in these reactions, yet I have not seen this problem mentioned in the literature, nor did my physician or genetic counselor discuss it with me. I am writing now to bring it to the attention of clinicians with pregnant patients undergoing diagnostic amniocentesis. I suggest to both clinicians and their patients that, when weighing the relative risk and benefits of prenatal diagnosis performed later (amniocentesis) as compared with earlier (chorionic villus biopsy), they not underestimate the negative effects of a 4½/ month wait before the woman knows if she is “really” pregnant.”

The next day, before I had mailed this letter, I received the results, and unfortunately they were the dreaded ones: trisomy 21. I have since then had the grim second-trimester abortion. From my current perspective of grief and shock, I encourage clinicians to help their patients avoid the denial described in my letter. My husband and I spared ourselves no pain by holding back emotionally. It has become a cultural expectation that one will keep one's pregnancy a secret until one has had the “all clear” from the amnio. One reasons, “If we get a bad result, we won't have to tell anyone.” But I now believe that reasoning is wrong. After our bad result, my husband and I did tell everyone. Sympathy and support from our friends, family, and colleagues have helped us to survive the ordeal of aborting a wanted pregnancy. By keeping the loss a secret, we would have cut ourselves off from such support when the feared outcome did happen.”

Not every couple will react this way, some preferring to keep their personal affairs private, but many will. The counselor needs to acknowledge these criticisms, and to rise to the challenge of providing a sympathetic and skilful service to clients/patients, according to their varying responses to deciding to have, undergoing, and waiting for the results of prenatal diagnosis, and then to support those who do get an abnormal result. These issues are addressed in detail in Prenatal Diagnosis: The Human Side (Abramsky and Chapple, 1994).

A considerable fraction of pregnant women are, in any event, opposed to invasive prenatal testing. In a study of pregnant women (age 37 and older) who had not undergone PND in Victoria, Australia, 33% had actively declined, with the two main reasons being concern about the safety of the test, and a conviction that they would not in any event have a termination. Another 6% had never been offered testing, these being for the most part women from minority groups, with single women also overrepresented. It is a challenge to see that all who might wish to have the choice of PND are indeed given it (Halliday et al., 2001).