The use of preimplantation genetic diagnosis (PGD) to screen embryos for aneuploidy and genetic disease is growing. New uses of PGD have been reported in the past year for screening embryos for susceptibility to cancer, for late‐onset diseases, for HLA‐matching for existing children, and for gender. These extensions have raised questions about their ethical acceptability and the adequacy of regulatory structures to review new uses. This article describes current and likely future uses of PGD, and then analyses the ethical issues posed by new uses of PGD to screen embryos for susceptibility and late‐onset conditions, for HLA‐matching for tissue donation to an existing child, and for gender selection. It also addresses ethical issues that would arise in more speculative scenarios of selecting embryos for hearing ability or sexual orientation. The article concludes that except for sex selection of the first child, most current extensions of PGD are ethically acceptable, and provides a framework for evaluating future extensions for nonmedical purposes that are still speculative.
Key words: embryo screening/ethics/gender selection/HLA matching/preimplantation genetic diagnosis
Debate about new genetic and reproductive technologies has often cited preimplantation genetic diagnosis (PGD)—the technique by which early human embryos are genetically screened and then discarded or placed in the uterus—as a reproductive practice that needs close ethical, legal, and social scrutiny. Reports that embryos are being screened for new indications such as susceptibility conditions, late‐onset diseases, HLA matching for existing children, and gender, reinforce the need for greater public attention. The purpose of this article is to describe current and likely future uses of PGD, and analyse the ethical, legal, and policy issues which they pose.
State of the art
PGD has been available since 1990 for testing of aneuploidy in low prognosis infertility patients, and for single gene and X‐linked diseases in at‐risk couples. A report in July, 2001 on worldwide use of PGD since 1990 reported that embryo or polar body biopsy has occurred in more than 3000 clinical cycles, with a 24% pregnancy rate, which is comparable with assisted reproductive practices which do not involve biopsy (Conference Report, 2001). More than 1000 children have now been born after PGD, and many pregnancies are ongoing (Tanner, 2002). The available data also indicates a 97% accuracy rate in preimplantation diagnosis, despite a high rate of allelic drop‐out in single cell PCR and the possibility of mosaicism or chaos in tested embryos.
Although some 40 centres worldwide have done PGD, four centres (Chicago, St Barnabas, Bologna, and Brussels) accounted for 2774 of the 3000 reported PGD cycles, and had an overall 29% clinical pregnancy rate (Conference Report, 2001). All four centres have done most of their PGD for aneuploidy, but Brussels has also concentrated on single gene analysis of blastomeres, and Chicago on polar body analysis. Other centres in the United States and Western Europe also do PGD, including four centres in London and centres in the Eastern Mediterranean, in Southeast Asia, and in Australia. While expansion in the number of centres offering PGD is to be expected, the investment of time, skill, and equipment needed for a competent programme may limit the number of IVF programmes providing this procedure. More work assessing the science, safety, and efficacy of PGD is needed as the field develops.
The demand for PGD
Although the use of PGD will grow, only a small number of people using assisted reproduction, and thus a much smaller of people overall who reproduce, are likely to use PGD to screen embryos. One reason is that accessing embryos through IVF is intrusive and expensive, and for some people raises serious ethical concerns. Thus patients will only seek it if there is a sufficient reason for them to do so, and insurance coverage and access to PGD facilities is provided.
PGD is least burdensome for a couple already going through IVF for infertility due to advanced maternal age. The additional costs of embryo biopsy and karyotyping could be justified if they produce a marked improvement in pregnancy rates. Non‐IVF patients, however, would seek PGD only if they were at special risk for genetic disease, and wanted to have a child without risking an affected pregnancy or later abortion. In both cases prenatal diagnosis would still be done to confirm that a fetus does not have the condition that the couple is seeking to avoid.
Several new indications for PGD single gene mutational analysis have been reported. These indications should increase demand for PGD from some couples unwilling to use prenatal diagnosis and selective abortion for those conditions. New uses include PGD to detect mutations for susceptibility to cancer and for late‐onset disorders such as Alzheimer’s diseases. In addition, parents with children needing haematopoietic stem‐cell transplants have also used PGD to ensure that their next child is free of disease and is a good tissue match for an existing child. Some people are also requesting PGD for gender selection for both first and later born children.
The emergence of new uses for PGD have made PGD a frequent target of ethical commentary and speculation about a future of greatly increased genetic selection and manipulation of offspring (Fukuyama, 2002; Kass, 2002; Stock, 2002). However, the low penetration of PGD in reproductive practice is likely to continue for some time. Even if PGD becomes the standard of care for older IVF patients, it is unlikely because of cost and lack of need to spread to good prognosis patients. Its use for screening for susceptibility and late‐onset diseases is limited by the few diseases for which single gene predispositions are known. In addition, relatively few parents will face the awful dilemma of a sick child in need of a haematopoietic stem‐cell donor, that might lead them to consider PGD. Gender preferences in offspring, however, might lead to demand for PGD in some cultural contexts. In short, PGD is likely to be increasingly sought by patients in narrow risk groups and to some extent for sex selection, but that demand will remain a small percentage of people using assisted reproductive and prenatal services. Despite the relatively limited use of PGD, the ethical, legal, and policy issues that arise from these new uses require attention.
Ethical issues in current and expanded uses of PGD
Two main sets of ethical objections make PGD and proposals for its extension controversial. One set of objections arises from the need to create and then select embryos on chromosomal or genetic grounds, with the deselected embryos then usually discarded. Other objections concern the fact of selection itself.
Objections to PGD based on its effect on embryos replay debates over abortion and embryo status that have occurred in many other contexts, from abortion to embryonic stem cell research. People who think that the embryo or fetus is a person will object to creating and destroying embryos, and oppose most uses of PGD. Others believe that preimplantation embryos are too rudimentary in development to have interests or rights, but that they deserve special respect as the first stage toward a new person (American Society of Reproductive Medicine, 1994). Under this view PGD is ethically acceptable when done for good reasons, such as preventing offspring with serious genetic disease. Indeed, PGD may prevent selective abortions for those diseases. A major issue with new uses of PGD is whether they sufficiently benefit important human interests to meet the demands of special respect for embryos that supporters of PGD may require.
A second set of objections arises from the fact of selection itself, and the risks of greatly expanded future selection of embryos and children. Sometimes based on religious views about the nature of human reproduction, ethical objections to selecting offspring traits raises two kinds of ethical concerns. One kind is deontological—the ethical judgement that it is wrong to choose traits of offspring, no matter how well intentioned. Dr Leon Kass has articulated this view, as has the President’s Bioethics Council in the United States, which he chairs. They argue that human reproduction is a ‘gift’ and that any form of selection or manipulation turns the child into a ‘manufacture’ and thus impairs human flourishing (Kass, 1998, 2000, 2002; President’s Bioethics Council, 2002). The second kind of concern is consequentialist. It arises from fears that increasing the frequency and scope of genetic screening of prospective children will move us toward a eugenic world in which children are valued more for their genotype than for their inherent characteristics, eventually ushering in a world of ‘designer’ children in which genetic engineering of offspring becomes routine.
While recognizing the strong objections of some people to PGD on these grounds, the following discussion assumes that the use of PGD to screen for aneuploidy and serious Mendelian disorders is ethically and legally acceptable when performed according to applicable regulatory guidelines. It concentrates instead on new indications for PGD, and asks whether they would also meet ethical standards of acceptability.
Expanded uses of PGD in assisted reproduction
The most common use of PGD is to screen embryos in assisted reproduction for chromosomal abnormalities before transfer. At present there is no reliable way to assess the viability of in‐vitro embryos other than by visual examination of morphology, but that method is not reliable, especially in low prognosis patients >39 years of age.
Testing for aneuploidy is now limited by the practice of examining only a few chromosomes. Some centres are now introducing full karyotyping of single blastomeres or polar bodies, which could improve accuracy in selecting viable embryos for transfer. A much greater use of PGD for low prognosis patients may be expected. Indeed, some practitioners have suggested that aneuploidy analysis for this group may soon become the standard of care, though overall efficacy has not yet clearly been established (Conference Report, 2001).
PGD is also useful for couples with chromosomal translocations who have experienced repeated spontaneous abortion. Robertsonian translocations, for example, may occur as frequently as 1 in 1000 in many populations. Centres in New Jersey, Chicago, and London (Guy’s & St. Thomas’) have developed methods of identifying unbalanced translocations that enable couples with a history of spontaneous abortion or other problems to have children (Conference Report, 2001; Scriven et al., 2001).
The use of PGD to exclude aneuploid embryos from transfer raises few special ethical issues beyond the use of IVF itself and discard of embryos. Nontransfer of embryos on visual grounds is now standard practice, so improving the grounds for selection is a gain. Nor does karyotyping for aneuploidies or translocations affecting fetal viability implicate ethical concerns about selecting embryos on nonmedical grounds. Ethical problems could arise, however, if programmes unequipped to do these procedures offer them to patients. Careful patient selection is also important, for PGD is not always indicated, even for patients with genetic anomalies (Scriven et al., 2001). Consultation with institutional ethics and research committees is essential before a programme begins to offer PGD services.
PGD for Mendelian diseases
In addition to full or partial karyotyping of chromosomes, PGD also allows direct examination of specific sections of the genome for mutations. Close to 1000 PGD cycles have been done for single gene disorders, resulting in at least 200 clinical pregnancies and 130 children born to date. More than half of these were done by the Brussels and Chicago centres, with Brussels specializing in blastomere biopsy and Chicago in polar body analysis. (Conference Report, 2001). Mutational analysis requires additional skills beyond karyotyping, including the ability to conduct multiplex PCR with nested primers of the gene of interest and related markers.
PGD from its inception has been touted as an important alternative for couples who were carriers of autosomal recessive, dominant, or sex‐linked diseases such as Tay Sachs, cystic fibrosis, sickle cell anaemia, Duchenne muscular dystrophy, and haemophilia. Rather than take a chance that a child would be born affected, some couples would go childless, adopt, or use donor gametes. Some at‐risk couples have conceived and then undergone prenatal testing. If the result was positive, they then had the option of terminating the pregnancy, a difficult and burdensome decision.
PGD offers the alternative of screening embryos rather than fetuses, thus avoiding the need to terminate a pregnancy in order to ensure an unaffected child. A couple selecting this approach would have to forego coital conception and undergo the costs and rigours of IVF, but may have a higher chance of successful pregnancy than would most infertile couples using IVF. Chorion villus sampling or amniocentesis later in the pregnancy is still advised to confirm the negative diagnosis.
PGD to screen out embryos affected by Tay Sachs, sickle cell, cystic fibrosis or other serious diseases raises no ethical issues beyond those that arise with prenatal diagnosis and selective termination of pregnancy generally. Many at‐risk couples might find embryo selection preferable to pregnancy termination because of the undeveloped state of embryos compared with fetuses. Unless PGD were available, some at‐risk couples might forego reproduction rather than risk an affected child or terminating an affected pregnancy.
PGD for susceptibility conditions
A logical extension of PGD for Mendelian disorders is its use to avoid the birth of children who are healthy at birth but face a higher than average risk of having cancer or some other serious disease. Indeed, PGD has been carried out to avoid the birth of a child with P53 mutations—in this case the Li‐Fraumeni syndrome (Simpson, 2001; Verlinsky et al, 2001a), and may be sought for BRCA1 and 2 susceptibility for breast cancer. Ethically, the question is whether the burdens of carrying susceptibility genes is so great for the child and parents that the burdens of IVF and PGD to screen embryos to avoid the affected children are justified. Many couples with these conditions might choose to go childless, rather than subject their child to the monitoring and worries that having known susceptibility genes carry. PGD may now make it possible to establish pregnancies free of the feared susceptibility condition.
The arguments in favour of such a use are similar to the arguments for avoiding the birth of children with a serious autosomal or X‐linked condition. Parents have a strong, interest in having children who will be healthy and not face the burdens of continued monitoring, prophylactic surgery, or other preventive actions, none of which is guaranteed to prevent the disease. If one accepts that embryos lack rights and interests but deserve special respect, then a plausible case for permitting PGD for this purpose exists. Creating and destroying embryos to have a healthy child does not treat embryos in a cavalier or frivolous way, and thus is consistent with special respect due embryos. The fact that the disease state does not occur until much later in life, unlike other Mendelian disorders, should not be morally significant. Having a child with inherited susceptibility to cancer could be a major source of suffering for parents and child, and could well affect a parental decision to reproduce (Simpson, 2001).
PGD for susceptibility conditions is legal in the United States. The Human Fertilisation and Embryology Authority (HFEA) in the UK has not yet granted licences for PGD for susceptibility testing. But it would fit within the statutory framework and should be permitted when application is made.
The chief legal restriction on susceptibility testing at present may come from people holding patents on the genes tested for by PGD. Myriad Corp., which holds the patent on the BRCA1 and 2 genes, requires that all tests be done on samples shipped to Utah, and refuses to offer the test for reproductive purposes (Myriad Policy Brochure, 1998). If their patents are valid and are enforced, women with breast cancer genes may not be able to use PGD to screen out daughters with those mutations. But this is an artefact of patent law, which gives patent holders the right to stop even highly beneficial uses of their inventions, not the ethics of PGD. A challenge to the Myriad patents, now before the European Patent Office, could, if successful, increase access to BRCA gene testing by PGD in Europe.
PGD for late‐onset conditions
PGD has also been used by a woman who carried a gene for early onset Alzheimer’s disease (AD), and who wished to have a child that would be free of that condition. In that case the woman was 31 years of age, had an older sister who had already experienced early onset Alzheimer’s, and had herself tested positive. She requested PGD to be sure that any fetus that she carried did not also have that gene. PGD was carried out, and she gave birth to a child free of that condition (Verlinsky et al., 2002).
The ethical issues raised here are similar to those raised for susceptibility conditions with one additional feature. As in those cases, avoidance of a child who will be healthy for a number of years but then in her 30’s or 40’s will experience AD [or Huntington’s disease (HD)] is a substantial, non‐trivial reason for employing PGD, as weighty as its use to avoid a child with susceptibility or disease genes. Indeed, the case is even stronger because the diseases tested for are not preventable, as may be the case with susceptibility genes.
The ethical twist in this case, however, concerns the ability of the affected parent to raise the child. Because late onset conditions like AD and HD are dominant, people would not be tested unless they knew that they carry (or are at risk of carrying) the disease gene themselves and faced a greatly shortened life span. Despite this knowledge, they wish to have offspring, but only if they can be sure that their child will not also have the disease. But having a child free of the disease means that that child is likely to lose a parent to the disease while still dependent on them.
The ethical issue is whether physicians act properly when they enable a woman/couple to have a child knowing that the child may soon lose one parent. Some ethicists have questioned whether helping to have a child when the parent may die prematurely is a responsible act, and some physicians or programmes may refuse to offer such services (Towner and Loewy, 2002). Yet a plausible case exists that such actions are ethical. People in this situation have as strong an interest in reproduction as other people receiving infertility services. The fact of a shortened life span makes reproduction for that person all the more acute and pressing, and is the only chance for her partner to reproduce with her. To be consistent, denying reproductive services in this case would also require denying sperm preparation and intrauterine insemination when an individual or her partner has HIV, providing IVF to a woman in her 30’s with cystic fibrosis, or initiating a pregnancy with a cancer patient’s gametes or embryos that were stored prior to radiation or chemotherapy.
It is true in all these cases that the child faces, with varying degrees of risk, having one of her parents die earlier than generally occurs, and will suffer grief and loss as a result, but another parent or care‐giver is almost certain to be present. In addition, the child in question could not otherwise be born but with this risk. The psychological trauma of dealing with the death or illness of a parent does not make a life so full of suffering or without clear benefit that one is harming the child by enabling her birth. As a result, one cannot persuasively condemn parents (and the physicians who help them) who reproduce in those circumstances as causing unnecessary or undue suffering for their children.
PGD for HLA matching for an existing child
PGD has also been used to enable a family with a child with Fanconi anaemia to have another child who would serve as a source of haematopoietic stem cells obtained from that child’s umbilical cord blood (Verlinsky, 2001b). Without a stem cell transplant, the first child is likely to die. Non‐sibling matches, even if available, are not nearly as safe and effective as sibling donations. If no sibling exists, parents might conceive coitally and hope to have a child who is a suitable match. Although rare, some parents might undergo prenatal diagnosis and abort if the fetus is not a good match, or even carry the fetus to term and give up the child for adoption (Auerbach, 1994).
Parents willing to have another child to produce stem cells for an existing child will find PGD attractive because it enables them to transfer to the uterus only embryos free of the disease and HLA‐matched to the existing child. The situation differs from other situations discussed because the PGD is done wholly or in part to choose embryos that will enable the resulting child to serve as an HLA‐matched donor for an existing child. Yet this distinction should not matter morally (Pennings et al., 2002; Robertson et al., 2002). Although helping an existing child is a motivating factor in having an additional child, such a reason alone is not likely to make the parents ignore the needs of the new child. With stem cells available from umbilical cord blood no further intrusions on the new child will occur. Its birth might save the life of an existing, loved child, which would only increase its specialness. If the stem cell transplant fails, the parents will have a surviving child to love. Creating and discarding embryos to provide donor cells for an existing child is not a trivial or frivolous use of embryos, and thus is consistent with the special respect that the nonright‐to‐life position accords to human embryos.
The HFEA has found PGD for HLA matching to be acceptable where the children born after PGD were themselves at risk for the condition to be treated in the existing child (Human Fertilisation and Embryology Authority, 2002). This means that it would be acceptable to use PGD for HLA matching in cases of Fanconi’s anaemia, but not childhood leukaemia or lymphoma unconnected to a known genetic mutation that later children might inherit. This limitation might be a prudent step to ensure that expanded uses of PGD occur slowly, but may not be entirely justifiable in terms of the principles that underlie allowing PGD for Mendelian disorders or other conditions. Whether or not the embryo itself is at risk for the disease for which stem cells are sought, screening for closeness of HLA match for an existing sick child serves a substantial family need and is not abusive or commodifying of resulting children (Robertson et al., 2002). The HFEA limitation should be reconsidered, and should not be followed in other jurisdictions (facing the issue).
PGD for gender selection
A controversial use of PGD is for nonmedical sex selection—to serve parental interests in having a healthy child of a particular gender. Because PGD for gender selection requires karyotyping only the sex chromosomes, it is more easily done than karyotyping for other aneuploidies or than single gene mutational analysis.
Requests for PGD for gender selection have come from two different groups. One is from people who wish to select the sex of their first born child. In almost all cases the first‐born preference is for a male child due to cultural mores that prefer males to perform certain rituals or to carry on the family line, or from rank sexism (Robertson, 2001). The second group is from people who already have a child of one gender and wish to have a child of the opposite gender. In many cases those requests are made after a family has had two or more children of the same gender, with no greater preference for males than for females.
In either case PGD for nonmedical gender selection is highly controversial, because gender selection itself seems to be a less compelling reason for selection of embryos for transfer. If sex selection for first children were carried out on a large scale, it could lead to great disparities in the sex ratio of the population, as has occurred in China and India (Sen, 1990; Eckholm, 2002). Since PGD is unlikely to be widely practised, however, its use for the first child is only marginally likely to contribute to those disparities. However, its use for first children is likely to reflect culturally‐founded sexist notions. As middle and upper classes in those cultures expand and have the means to obtain PGD, such demand could grow.
The use of PGD to select the sex of second or subsequent children is much less susceptible to a charge of sexism if used to choose a gender opposite to that of an existing child or children. Here a couple seeks variety or ‘balance’ in the gender of offspring because of the different rearing experiences that come with rearing children of different genders. Psychologists now recognize many biologically‐based differences between male and female children, including different patterns of aggression, learning, and spatial recognition, as well as hormonal differences (Robertson, 2001). It may not be per se sexist to wish to have a child or children of each gender, particularly if one has two or more children of the same sex.
Some feminists, however, would argue that any attention to the gender of offspring is inherently sexist, particularly when social attitudes and expectations play such an important role in constructing sex role expectations and behaviours. Other feminists find the choice of a child with a gender different from existing children to be morally defensible as long as “the intention and consequences of the practice are not sexist” (Mahowald, 2000). As United States Supreme Court Justice and former feminist lawyer Ruth Bader Ginsburg noted in an important sex discrimination case, “the two sexes are not fungible... inherent differences between men and women, we have come to appreciate, remain cause for celebration” (United States Supreme Court, 1996). Desiring the different rearing experiences that one has with boys and girls does not mean that the parents, who have already had children of one sex, are sexists or likely to devalue one or the other sex (American Society of Reproductive Medicine, 2001).
Based on this analysis the case is not strong for allowing PGD for the first child, but may be acceptable for gender variety in a family. With regard to the first child, one may be promoting or entrenching sexist social mores. One could also argue that creating embryos to choose the gender of the first child is not a strong enough reason to show special respect for embryos. A proponent of sex selection, however, might argue that in certain cultures having a first‐born male is so important that it is a legitimate reason for creating and destroying embryos. If PGD is not permitted, pregnancy and abortion might occur instead.
The case for PGD for gender variety is stronger because the risk of sexism is lessened. A couple would be selecting the sex of a second or subsequent children for variety in rearing experiences, and not out of a belief that one sex is privileged over another. Sex selection in that case would occur without running the risks of entrenching sexism and hurting women (American Society of Reproductive Medicine, 2001). But is the desire for gender variety in children, even if not inherently sexist, a strong enough reason to justify creating and discarding embryos on that basis? The answer depends on how strong an interest that is. Aside from anecdotal evidence, it is unclear how many parents are in that situation and how strong the desire for gender variety is. If it is a strong one, e.g., would stop parents from having another child unless they could be certain it would be of the sex opposite of existing children, then perhaps it should be acceptable as well. However, more evidence of the strength of the need would help in reaching a final conclusion (Robertson, 2002).
It is instructive to see how the American Society of Reproductive Medicine (ASRM), a leading American organization of fertility doctors, has dealt with this question. The Ethics Committee of the ASRM initially addressed the issue of PGD for sex selection generally, and found that it “should be discouraged” for couples not going through IVF, and “not encouraged” for couples who were, but made no distinction between PGD for selecting the sex of first and subsequent children (American Society of Reproductive Medicine, 1999). Subsequently, it found that preconception gender selection would be acceptable for purposes of gender variety but not for the first child (American Society of Reproductive Medicine, 2001).
Perceiving these two positions to be inconsistent, a doctor who wanted to offer PGD for sex selection inquired of the Ethics Committee why preconception methods for gender variety, which lacked 100% certainty, were acceptable but PGD, which guaranteed that certainty, was not. Focusing only on the sexism and gender discrimination issue, the Chair of the Ethics Committee, in a widely publicized letter, found that PGD for gender balancing would be acceptable (Kolata, 2001). When the full committee considered the matter, it concluded that it had not yet received enough evidence that the need for gender variety was so important in families that it justified creating and destroying embryos for that purpose (Robertson, 2002). In the future if such evidence was forthcoming then PGD for gender variety might also be acceptable.
What might constitute such evidence? One source would be families with two or more children of one sex who would have another child only if they could be sure that it would be a child of the sex opposite of existing children. Given the legitimacy of wanting to raise children of both genders, reasonable people might find that this need outweighs the symbolic costs of creating and destroying embryos for that purpose. Selection in that case could enable a couple to have a child that they would not otherwise have had, a goal served in other cases of embryo screening. A variation on this need might come from families that had suffered the death of a child and wanted to have another child of that sex. Still another variation has arisen in Israel, where the Ministry of Health has approved use of PGD to select a girl for a couple using an anonymous sperm‐donor, to prevent the disclosure of the male infertility that would likely occur if the child were male (Traubman and Shadmi, 2002).
Another arguably acceptable use of PGD for gender variety has been reported in India, where an IVF programme in Bombay is now providing PGD to select male offspring as the second child of couples who have already had a daughter (Malpani et al., 2002). Because of the importance of a male heir in India, those couples might have resorted to abortion if pregnant with a female fetus (even though illegal for that purpose). In that setting PGD for sex selection for gender variety appears to be justified.
PGD for nonmedical traits
Other controversial uses of PGD would arise if genetic tests for nonmedical traits such as hearing, sexual orientation, height, beauty, intelligence, or other factors became available (Kass, 1998; Fukuyama, 2002; Stock, 2002). Untangling the inherited basis of phenotypic traits strongly influenced by learning and environment will not be simple, and such tests, with a few exceptions, are unlikely to be available, if at all, for some time to come.
One potential would be tests for GJB2 mutations, which are the largest known contributor to inherited deafness (Nance and Pandya, 2002). If such a test is available, people with a family history of deafness might request PGD to screen out embryos with the mutation, in order to increase their chances of having a hearing child. Because hearing is clearly beneficial to a child, a major ethical concern with this practice would be the prejudice to the deaf community that it might cause. For such screening to be acceptable, it is essential that it not be offered in ways that denigrated the deaf or lessened respect for them.
The reverse situation—a deaf couple wants PGD to screen out embryos that do not have the GJB2 mutations—could also arise. Here the ethical issue is whether having a deaf instead of a hearing child would have hurt that child. A similar issue has arisen in a widely publicized case of a deaf lesbian couple choosing a deaf sperm‐donor in order to increase the chances of having a deaf child. Given the deaf parents’ strong commitment to the well‐being of their child and the rich culture now available to deaf people, many bioethicists analysing this case find the deaf parents’ choice to be acceptable (Davis, 2001; Fletcher 2002; Levy, 2002). Others, however, strongly disagree (Anstey, 2002).
Similar issues would arise if a genetic test for sexual orientation became available. A main ethical concern would be whether use of the test to avoid offspring with homosexual orientation would increase discrimination or antipathy toward gays. The concern with using the test to ensure a homosexual orientation in one’s child is whether this shows a responsible regard for the well being of the child. Much would depend on the precise circumstances of use and the characteristics of people seeking the test.
Some people would strongly urge that selection for hearing ability, sexual orientation, or other traits should never be permissible (Fukuyama, 2002; Kass, 2002). Reinforcing their concern is the fear that such uses of PGD will inevitably lead to more drastic efforts at selection and alteration of offspring traits. Although PGD acts negatively by screening out embryos, it accepts the principle that parents might rightfully exercise control over the genomes of offspring. It will then be much easier to take the next steps leading to positive alteration and engineering of offspring traits and the risk of ‘designer’ children which that will bring.
Although it is useful to identify the ethical issues that such uses would potentially raise, it is too soon to reach definitive judgements about whether these uses, if ever feasible, would or should be permitted. Such tests are too far off in the future to make informed judgements about them now. Until they are closer to practical reality, they should not be an important factor in determining the acceptability of more feasible uses.
Making public policy for PGD
Legitimate concerns about potential misuse of embryo screening and selection make it essential that a sustained public debate about these issues occurs as technical progress continues. Some of the discomfort that surrounds new uses of PGD stems from a sense in many countries that there is no effective oversight of its development and use.
The case is different in the UK, where the HFEA has legal authority over whether a clinic is licensed to do PGD at all and for what indications. Additional uses of PGD may occur only if the HFEA is satisfied that the uses are within statutory guidelines and the programme is qualified to undertake the work. In addition, the HFEA uses a public consultation process to assess public attitudes and draw up guidelines for new uses, as demonstrated in its consultation process for whether PGD was acceptable for HLA matching of embryos with existing sick children (Human Fertilisation and Embryology Authority, 2002). While some physicians have challenged the scope of the HFEA’s regulatory authority, the HFEA has provided a regulatory model that other nations might emulate.
The situation in the United States is quite different. No agency exists at the state or federal level that plays a role comparable with that of the HFEA. Congress exercises some control by refusing to fund research or use of PGD, but this often means that the activity escapes meaningful external review in the private sector. A few States have laws restricting embryo research, but several of them have been struck down as unconstitutionally vague or intrusive on reproductive rights, and the others are vulnerable to the same charges. State malpractice and tort law will apply, but those restrictions provide little oversight of the ethical acceptability of new procedures. How PGD is used and for what indications is thus left largely to the discretion of providers offering those services and the patients who seek it.
Even if there is no HFEA‐like authority for public oversight, extensive public debate is desirable as new techniques become available. With PGD, the need for a new agency or more explicit public controls may be less pressing than initially thought. As this article has discussed, most of the new uses of PGD now under consideration fit easily into old categories of efforts to ensure that offspring are healthy. More controversial uses, such as for gender selection, are not as readily accepted and practised, though some increase of PGD for gender selection, especially for gender variety, is likely. Except for deafness and possibly sexual orientation, using PGD to select for other nonmedical traits, if ever possible, is too far off in the future to justify final judgements about them now.
PGD is increasingly available for aneuploidy in low prognosis IVF patients and for single gene mutations that cause genetic disease, susceptibility to cancer, and late onset disorders. If PGD is acceptable to prevent offspring with serious genetic disease, then these additional uses should be acceptable as well. There is also ethical support for using PGD to assure that a child is an HLA match with an existing child. More controversial is the use of PGD for gender selection, particularly for the first child. Equally controversial would be its use to screen embryos for hearing, sexual orientation, and other nonmedical traits—uses that are now highly speculative. Careful ethical analysis and open public debate is essential if new uses of PGD are to become acceptable methods for having children.
Support was received from the US National Institutes of Health, National Institute of Human Genome Research, Ethical, Legal, and Social Issues Program.
1999) Ethics Committee, Sex selection and preimplantation genetic diagnosis.
1994) Ethics Committee, Ethical considerations of assisted reproductive technologies.
2001) Ethics Committee, Preconception gender selection for nonmedical reasons.
2002) Are attempts to have impaired children justifiable?
J. Med. Ethics,
1994) Umbilical cord blood transplants for genetic disease: diagnostic and ethical issues in fetal studies.
Reprod. Biomed. Online,
2001) Genetic Dilemmas: Reproductive Technology, Parental Choices, and Children’s Futures, Routledge, New York, USA
2002) Desire for sons drives use of prenatal scans in China. New York Times, June 21, 2002, p. A3.
4.1. Single Gene Disorders
PCR has been the preferred method of diagnosis in PGD for single gene disorders. Nearly all genetically inherited single gene disorders where sequence information on the mutation is known can be detected by PGD. Diagnostic protocols for over 200 different single gene disorders have been reported, with the most common being for cystic fibrosis, spinal muscular atrophy, hemoglobin disorders, such as β-thalassemia, Huntington disease, fragile X syndrome and myotonic dystrophy. As opposed to the analysis of DNA from blood or amniocentesis/CVS samples, PCR in PGD usually begins from very limited amounts of DNA found in blastomere or trophectoderm biopsies. Single cells contain approximately six picograms of DNA, and for many PCR tests, 250 nanograms or more of DNA is required as the starting material. This implies that roughly a 40,000-fold amplification in DNA concentration is needed. One problem that became apparent was that the large number of PCR cycles needed to generate a product led to an increasing rate of the introduction of errors in the DNA. This was overcome by the development of “nested” PCR by Holding and Monk . This technique uses a first round of amplification with an outer set of primers, followed by a second round using a set of inner primers, which when using the first round product as a template, amplifies a smaller PCR product inside the outer primers, which is robustly produced with high accuracy.
The low initial starting concentration of DNA in embryo samples predisposes PCR to several other complicating factors, such as contamination with external DNA and allele dropout (ADO). Contaminating DNA can originate from the technician performing the PCR, from carryover of other PCR products or controls from earlier amplifications or from cumulus cells or sperm adherent to the zona. Careful cleaning of the embryo prior to fertilization and the use of intra-cytoplasmic sperm injection (ISCI) of a single sperm for oocyte fertilization can avoid this latter type of contamination. Operators typically dress in surgical gowns and perform the PCRs in laminar flow hoods situated in “clean rooms” to limit external DNA contamination.
Allele drop out (ADO) refers to preferential amplification or failure of the amplification of one of the alleles present in a heterozygous sample. This could lead to misdiagnosis of a heterozygous embryo as a homozygous affected sample, in which case, a normal embryo would be excluded from the cohort of embryos available for transfer. In the case of a dominant genetic disease, ADO can cause an abnormal embryo to be misdiagnosed and transferred in error as “normal”. ADO can occur if the annealing of one of the two PCR primers is less efficient. Alternatively, inefficient cell lysis or degenerated DNA can theoretically lead to ADO. The frequency of ADO is variable and, in PCR of single blastomeres, can be seen in 5% to over 20% of amplifications [45,46].
The early evaluation of PCR products included the analysis of products of radioactive-labelled primers or nucleotides, as well as ethidium bromide staining of DNA in agarose or polyacrylamide gels after restriction enzyme digestion. These methods have been replaced by fluorescent labeled probes or nucleotides, yielding products detected and sized by capillary electrophoresis. The most frequently utilized method for mutation detection is DNA primer-extension minisequencing  in which a probe is developed that extends up one base prior to the first nucleotide of the gene alteration. Single base-pair changes or small deletions/duplications can then be detected by the extension of the primer with fluorescent-labeled dideoxynucleotides, which will identify the next base in the target sequence and then stop, due to the chain-termination effect of dideoxynucleotides. Commercially available kits, such as the ABI SNapShot™ protocol (Life Technologies, Carlsbad, CA, USA), can be used for this purpose. Deletions and duplications can also be detected by that sizing of a PCR product, including the region of interest. The PCR reactions include primers for multiple target sites, not only for the area containing the mutation under consideration, but for closely linked polymorphic short tandem repeat (STR) markers, which can be used to detect DNA contamination and ADO . This multiplex fluorescent PCR protocol led to a significant decrease in the number of analyses compromised by contamination or ADO.
A pretest work-up typically involves obtaining DNA samples from the at-risk couple, as well as an affected child or grandparents to allow phasing of the STR haplotype with the mutation being tested. This allows the determination of the two maternal and two paternal haplotypes, as shown in Figure 1. This analysis aids in the interpretation of the results of PGD testing, although recombination of STR markers with the genetic mutation can be seen.
Pre-PGD workup for a family with a previous child with spinal muscular atrophy. Panel (a) shows how the study of both parents and grandparents allows the phasing of the SMN mutation relative to polymorphic short tandem repeat (STR) markers; panel (b)...
Newer PCR protocols have been developed using whole genome amplification (WGA) as a universal first step after biopsy of a single blastomere or 5–9 cells from a trophectoderm biopsy. This allows the analysis of a number of different sequences, such as compound heterozygous mutations of the CFTR gene, as well as allowing parallel testing of single gene disorders and linked STR markers or aneuploidy testing by aCGH. There are a number of methods for WGA that have been developed. Early protocols included primer extension PCR , degenerative oligonucleotide primed (DOP) PCR  or blunt end ligation and PCR amplification (GenomePlex™, Rubicon Genomics, Ann Arbor, MI, USA) . All of these methods were dependent on PCR and carried the potential for uneven representation of the genome by preferential amplification or ADO. Multiple displacement amplification (MDA) is a WGA technology based on isothermal “rolling amplification” after the binding of random hexamers to the target sequence [50,51]. Despite the isothermal amplification in MDA protocols, ADO rates have been found to be at least 4% with SNPs , while others found a much higher ADO rate closer to 25% for linked STR markers used for PGD haplotyping [52,53]. The significant issues with ADO in single gene disorder PGD can be alleviated by the inclusion of multiple cells in the biopsy sample, which is why many centers are now moving toward trophectoderm biopsy, where 5–10 cells are typically obtained. In our lab, using trophectoderm biopsy and GenomePlex™ amplification, we have encountered an ADO rate of approximately 5% in single gene disorder PGD. Zong et al.  have reported on a new WGA method of multiple annealing and looping-based amplification (MALBAC), which is reported to significantly reduce amplification bias and provide 85%–95% genome coverage with 25X sequencing depth compared to 72% coverage at the same sequencing depth for the MDA protocol.
The GenomePlex™ system presents another issue, in that the average size of the amplified product is 500–1000 bp. In using this DNA for single gene disorder testing, we have found it important to limit the amplicon size, such that PCR products are smaller than the average WGA DNA product length and, therefore, contained in a single fragment. The use of larger amplicons leads to the failure of the PCR to include the desired sequence. Tran et al.  showed that this approach is successful in developing PGD tests and analyzing results for different classes of mutations, including single base pair changes, deletions/duplications and triplet repeat expansion mutations along with STR repeats, all with a low ADO rate.
It is likely that all WGA protocols result in some biased amplification, leading to decreased genomic coverage. Newer products are now being developed to address this bias, as well as producing highly accurate and representative amplification. As interest in more extensive sequencing in PGD develops, these issues will become more critical; however, at the moment, more limited evaluation, such as single gene sequencing and shotgun sequencing-chromosome counting protocols, seems to produce acceptable results with some current WGA methods (see below).
4.2. Age-Related Chromosome Aneuploidy
The most common use of PGD involves the analysis of embryos for chromosome aneuploidy arising from either meiotic mal-segregation in oocytes or (less commonly) sperm or mitotic abnormalities of embryos after syngamy, which results in mosaicism; defined as the presence of two or more cytogenetically defined cell lines in the same embryo. Chromosome aneuploidy is strongly correlated with maternal age, however; mitotic abnormalities seem to be age-independent . Embryos that are aneuploid often result in developmental arrest, failed implantation or subsequent miscarriage . Chromosome analysis of miscarriage products by aCGH shows that approximately 70% of first trimester miscarriage is due to aneuploidy [58,59]. Since chromosome anomalies are the major cause of unsuccessful IVF or pregnancy loss, it was proposed that identification and preferential transfer of euploid embryos in IVF would lead to higher implantation and birth rates with a lower miscarriage rate [9,60]. This concept was felt to make perfect sense, and it was initially expected that aneuploidy testing would result in a major improvement in IVF outcome, especially in couples with recurrent pregnancy loss or recurrent IVF failure. Since standard metaphase karyotyping cannot be rapidly performed on polar bodies, blastomeres or trophectoderm cells, a different technology of interphase FISH was introduced to evaluate the chromosomal content of embryonic cells. Clinically, FISH was first performed for sexing, using X and Y probes to avoid the transfer of embryos with X-linked genetic disease. Shortly thereafter, more extensive chromosome testing was developed using protocols that score 5–12 chromosomes in multiple rounds of hybridization using centromeric or locus-specific FISH probes [7,8,9]. Testing was limited by the fact that there were only five visible spectrum fluorochromes available to label FISH probes, so testing needed to be performed in multiple sequential rounds of hybridization, stripping and re-hybridization; each round performed with decreasing efficiency.
This form of PGD, which did not involve genetic disease testing and was designed to improve the outcome of IVF, was termed preimplantation genetic screening (PGS). The term was first used in Europe, apparently to distinguish chromosome testing from single gene disorder analysis, which was considered a less controversial use of PGD. The name PGS is a misnomer, as aneuploidy testing by PGD is clearly not a screening test in that clinical action (i.e., the non-transfer of embryos diagnosed as abnormal) is taken based on the results of the testing. This is in contrast with the classical definition of screening tests, which alters one’s odds of having a specific condition without providing a definitive, actionable diagnosis. An example of a screening test is first trimester combined screening (ultrasound and blood analyte measurement), which does not provide definitive chromosomal analysis, but adjusts a patient’s prior risk based on maternal age alone.
Most chromosome studies performed prior to 2007 on cleavage stage embryo biopsy samples utilized FISH for analysis and showed significant levels of mosaicism throughout preimplantation development [36,61]. A systematic review of chromosomal mosaicism in preimplantation embryos, where every cell of an un-transferred embryo was analyzed , showed that only 22% of embryos were uniformly diploid, 73% were mosaic and 5% had multiple non-mosaic abnormalities. Of those that were mosaic, nearly half had diploid/aneuploid mosaicism. Comparative genomic hybridization (CGH) is an alternative method of chromosome analysis originally developed to karyotype solid tumors. DNA from a test sample and control sample are labeled with either green or red fluorochromes. The DNAs are both hybridized in equal concentration to either metaphase chromosomes (mCGH) or to chromosome specific DNA probes fixed to a microscope slide (aCGH) . The fluorescent intensity of the two fluorochromes is measured; if equal, a yellow color is obtained. A red or green color is seen if either the test or control DNA is present in excess. A commercial aCGH protocol (24Sure™ Bluegnome, Cambridge, UK) has been developed specifically for PGD, which contains about 3600 very robust BAC clones (average size of 150 kb) spaced at approximately 1 MB intervals with 25% genomic coverage and a resolution of 4–10 MB. The protocol can be completed in as little as 12 h, allowing a fresh Day 6 transfer after Day 5 trophectoderm biopsy. The results are presented as log2 ratios, as shown in Figure 2.
Array comparative genomic hybridization (aCGH) tracing after trophectoderm biopsy: (a) normal male embryo (female embryo control in blue); (b) female embryo with monosomy for chromosome 20 (male control in red); (c) an excellent quality blastocyst showing...
Early studies using mCGH of all blastomeres of non-transferred embryos [48,63] revealed that only 25% of the Day 3 embryos were totally euploid and that as high as two thirds demonstrated mosaicism. In one-half of these embryos, there were no normal cells (aneuploid/aneuploid mosaic), and 25% were diploid/aneuploid mosaics. These results led many to question whether the analysis of a single Day 3 blastomere was a good indicator of the developmental potential of the embryo.
Early data using aneuploidy testing with FISH in women of advanced reproductive age seemed to indicate a trend toward improvement in implantation and pregnancy rates [64,65,66,67,68]. These studies were mostly small, case-control studies in patients with advanced reproductive age or recurrent pregnancy loss. Despite somewhat encouraging results, statistical power could not be generated, and therefore, it was recommended that randomized clinical trials (RCT) be undertaken to prove that aneuploidy testing via PGD led to an increase in the birth of healthy children. Staessen et al.  in 2004 reported the first RCT comparing the results of blastocyst transfer with or without PGD using a FISH panel of seven chromosomes after the Day 3 cleavage stage biopsy of two cells. The results showed improved implantation and clinical pregnancy in the PGD group, but the difference was felt to not be of statistical significance. This group later performed a second RCT employing only single cell embryo biopsy, which produced improved results compared to the biopsy of two cells, but did not demonstrate any statistically significant benefit for the implantation or ongoing pregnancy in the PGD group, as compared to controls without PGD . Overall, 10 RCTs have been performed to evaluate the efficacy of aneuploidy testing (for a summary table, see Fragouli and Wells ). The 2007 multicenter double-blind RCT performed by Mastenbroek et al.  reported that Day 3 blastomere PGD actually led to a decrease in the ongoing pregnancy rate, with the PGD group having a 25% ongoing pregnancy rate, as compared to 37% for the control group. This study garnered much media attention and caused much alarm among patients contemplating chromosome-based PGD. It was pointed out that the study was flawed on multiple levels, from patient selection, lack of embryo biopsy experience, selection of incorrect chromosomes for study and a high incidence of no results [71,72]. Some investigators pointed out technical issues with the RCTs already performed and felt that strict adherence to their protocol was necessary to show improved pregnancy rates with FISH .
At this point in time, there was a re-evaluation of the technology of the cleavage stage PGD for chromosomal aneuploidy testing. Relevant factors included: (1) The reduction in the implantation of embryos after blastomere biopsy; (2) The experience of the embryology laboratory with blastomere biopsy; (3) Proper cell fixation and cell lysis technique; (4) Hybridization issues, such as failure, split or overlapping signals, or nuclear damage with micronuclei formation; (5) The selection of optimal probes for FISH analysis (the use of seven probes for chromosome X, Y, 13, 16, 18, 21 and 22 was shown to account for 72% of chromosome abnormalities seen in spontaneous miscarriage ); and (6) The presence of chromosomal mosaicism, which leads to decreased accuracy and poorer pregnancy outcome [75,76].
The clinical utility of FISH-based PGD for chromosome abnormalities was seriously questioned, and it was recognized that in order to keep recommending PGD for patients with advanced reproductive age, recurrent miscarriage or multiple failed IVF cycles, a new RCT was needed using a different source of embryonic tissue for analysis [77,78]. In Europe, the analysis of polar bodies was selected, due to the legal restrictions on IVF and PGD present in many countries. Other groups turned their attention to the blastocyst and trophectoderm biopsy. At the same time, researchers in the field again considered CGH for chromosome analysis. As noted above, in 2000, mCGH was used by Wells and Delhanty  and Voullaire and colleagues  to analyze DNA obtained from multiple blastomeres of untransferred embryos after degenerate oligonucleotide primed (DOP) PCR. These studies indicated that 51% of cleavage stage embryos were aneuploid in every cell, and 24% contained a mixture of normal and abnormal cells, with only 25% being uniformly normal. Although superior in detecting chromosome anomalies compared to FISH with limited panels of chromosomes, mCGH is very labor and time intensive, thus requiring all embryos to be cryopreserved with a later frozen embryo transfer using embryos determined to be normal. Originally, this protocol was complicated by the fact that cryopreservation of biopsied embryos was very poor with low survival . This improved dramatically with the introduction of embryo vitrification, with survival rates now over 90% and implantation similar to that of fresh embryo transfer .
Renewed interest in CGH resulted in this technology being validated for use in polar bodies [81,82,83], blastomeres and blastocysts [84,85,86]. Analytic improvement was accomplished with the evolution of mCGH to aCGH, where hybridization was performed on chips containing thousands of DNA segments fixed on a glass slide. Gutierrez-Mateo et al.  studied single blastomeres by aCGH after WGA in non-transferred embryos and compared the results to those obtained by FISH. They showed that aCGH detected 42% more chromosome defects, as compared to a panel of 12 FISH probes. Fragouli and colleagues  investigated the chromosomal status of blastocysts with mCGH. Some of these embryos had an abnormal Day 3 FISH result, which was confirmed by CGH, while the majority were donated, untested embryos, which were felt to represent what might be obtained in clinical practice. The aneuploidy rate of the blastocysts was 38.8%, which was lower than the 51% seen in Day 3 embryos [48,63]. Aneuploidy of all chromosomes was seen in blastocysts, including those of the larger chromosomes, as well as chaotic abnormalities (see Figure 2c). These findings dispelled previous beliefs that culturing to the blastocyst stage would effectively select against embryos with chromosome anomalies. The authors also showed the concordance of chromosome errors in the ICM and TE, indicating that preferential allocation of abnormal cells away from the ICM to the TE is not seen in humans.
In addition to mCGH and aCGH, single nucleotide polymorphism (SNP) arrays were validated to perform 24 chromosome analysis. SNPs are sequence variants that exist in the population, where at a particular locus, one of two (or more) nucleotides is commonly found. Biallelic SNPs, where one of two bases is seen, have proven to be useful in molecular cytogenetic analysis. The relative intensity of the two alleles, typically referred to as A and B, can be measured at heterozygous loci to detect duplications or deletions of chromosomal regions. SNP arrays also can reveal the parental origin of chromosomal errors by genotyping the parents. Rabinowitz and colleagues  have developed a bioinformatics algorithm using parental genotypes and reported on the chromosome number analysis of blastomeres from cleavage stage embryos. This showed increasing aneuploidy with maternal age originating from meiotic nondisjunction of the oocyte. SNP arrays also have the advantage of being able to detect polyploidy and uniparental disomy (UPD), which cannot be seen with CGH. Treff and colleagues  validated a high-density array with 262,000 SNPs using 99 WGA amplified single cells from chromosomally-defined lymphoblastoid lines. Three hundred thirty five WGA blastomeres were also studied with only 37.3% being euploid. This group also compared the performance of SNP arrays with that of interphase FISH by examining 160 blastomeres from arrested embryos and concluded that the SNP array was more accurate and FISH overestimated the aneuploidy rate . Finally, CGH and SNP arrays demonstrated the significant mosaicism of cleavage stage embryos, which could lead to misdiagnosis and the non-transfer of potentially viable embryos.
Biopsy of embryos at the blastocyst stage was considered after the development of sequential media that allowed in vitro culture to Day 5 or 6 after fertilization. These embryos were felt to have higher developmental potential, and the testing of trophectoderm cells could lead to improved clinical outcome with less damage to the embryo. Fragouli and colleagues  studied 64 non-transferred blastocysts by mCGH, aCGH and a nine-probe FISH panel and showed that 42% of blastocysts were uniformly euploid, 30% uniformly aneuploidy, 17% were diploid-aneuploid mosaics and 15.4% composed of different aneuploid cells. FISH results were concordant with CGH in 10 of 12 embryos; however, in the remaining two, only the sex chromosomes were consistent. Northrup and colleagues  used SNP arrays to study 50 blastocysts and demonstrated that 62% of the blastocysts were uniformly euploid, 7.8% uniformly aneuploid and the remaining 30% mosaic, containing either diploid/aneuploid or aneuploid/aneuploid mosaicism.
Studies evaluating the ability of aCGH to detect mosaicism in blastocysts have been reported by Mamas et al.  and Novik et al. . These studies created models of mosaic blastocysts by either mixing cells or DNA from euploid and aneuploid cell lines or embryos confirmed to be uniformly monosomic by FISH. The results showed that aCGH using the 24Sure protocol was able to clearly detect mosaicism when the proportion of abnormal cells was approximately 25%–30%.
As trophectoderm biopsy appeared to produce a more accurate assessment of the chromosomal status of the embryo, clinical studies were undertaken to evaluate the outcome of array-based PGD in infertile couples and those with recurrent pregnancy loss or repeated IVF failure. Schoolcraft and colleagues  studied 45 couples with infertility that underwent CGH-based PGD with trophectoderm biopsy and compared the results to 113 couples who had blastocyst transfer without PGD. Nearly 70% of the PGD couples achieved a clinical pregnancy as compared to 45% in couples where PGD was not performed. A clinical study in infertile couples using trophectoderm biopsy, SNP microarray analysis, embryo vitrification and subsequent frozen embryo transfer (FET) was reported by Schoolcraft and colleagues . In the first 100 FETs, the biochemical pregnancy rate was 87% with a 73% clinical pregnancy rate (positive fetal heart rate) and a 2.3% miscarriage rate.
To avoid the need for embryo vitrification and later FET, Treff and colleagues  developed a new method for 24 chromosome analysis of blastocysts using real-time quantitative PCR (qPCR). This method used multiplex PCR in a 384-well plate format to amplify two areas on each arm of all chromosomes. The analysis can be completed in 4 h, allowing for fresh embryo transfer. Each of the methods for 24 chromosome analysis described above has their own specific advantages and drawbacks. An excellent summary comparing the different technologies has been published by Handyside .
The results of early and ongoing clinical trials of blastocyst biopsy and 24 chromosome analysis on improving implantation and ongoing pregnancy have been encouraging . Harton et al.  have collected data from multiple IVF centers on patients undergoing IVF with PGD by aCGH after cleavage stage or trophectoderm biopsy. The pregnancy rates were higher for trophectoderm biopsy, but for both, the transfer of euploid embryos blunted the typical decrease in implantation and ongoing pregnancy associated with advancing maternal age. In other words, if euploid embryos were identified, their implantation potential was relatively independent of maternal age. This is what would be expected if aneuploidy was the major cause of implantation failure and miscarriage. Of course, as maternal age increased, so did the number of women who did not produce any euploid embryos. Scott et al.  have reported on an RCT involving 155 couples who were randomized to receive a Day 5 blastocyst transfer without PGD (control group) or a trophectoderm biopsy, comprehensive chromosomal screening by four-hour qPCR and Day 6 embryo transfer (study group). The rate of implantation and delivery was 66.4% with PGD and 47.9% in the control group. Delivery per cycle was 84.7% for the PGD group versus 67.5% in the control group.
These results also highlight one of the most important benefits of the increased implantation and lower miscarriage rate after 24 chromosome analysis; namely, the ability to optimize elective single embryos transfer (eSET). The goal is to reduce the incidence of twin gestation, which is associated with a 5–10-fold increase in the risk for fetal and maternal complications, including gestational diabetes, preeclampsia, premature delivery and low birth weight . Since embryos that have a normal chromosome analysis have the highest developmental potential, it makes sense that the transfer of euploid blastocysts will optimize eSET. An RCT for eSET in good prognosis women <35 years of age was carried out by Yang et al. , who compared the transfer of a single embryo at the blastocyst stage on Day 6 that was either tested by aCGH or assessed by morphologic criteria only. In the aCGH group, the ongoing (>20 gestational weeks) pregnancy rate was 69.1% compared to 41.7% in the morphology only group. Another RCT was carried out by Forman et al.  in 205 women <43 years of age, but with normal ovarian reserve testing. The study group had chromosome testing by qPCR with subsequent eSET, while the control group transferred their two morphologically best embryos. The ongoing pregnancy rate per patient was similar (60.7% in the eSET group versus 65.1% after the two-embryo non-aCGH tested transfer); however, the multiple pregnancy rate was significantly higher in the two embryo transfer group (53.4%) versus the PGD eSET group (0%). The conclusion was that 24 chromosome screening after trophectoderm biopsy and eSET produced as high a pregnancy rate as the transfer of two untested embryos without encountering the increased risk for twin gestations.
4.3. Structural Chromosome Rearrangements
Structural chromosome rearrangements include reciprocal and Robertsonian translocations and inversions. These can be seen in approximately 1/500 live-born infants and 1/250 prenatal samples . Individuals who carry balanced chromosome translocations (or inversions) generally have no clinical findings related to the translocation, but will produce high rates of abnormal gametes after meiotic segregation. The involved chromosomes will orient at the metaphase plate in a quadrivalent pattern and will segregate to the two daughter cells in one of 30 or so segregation patterns. These can contain areas of chromosome deletion or duplication, which can lead to pregnancy loss, failed implantation, apparent infertility or the birth of a child with physical and/or developmental disability.
Scriven and colleagues  published a method for the evaluation of segregation in translocations involving FISH using commercially available specific centromeric and sub-telomeric probes. Although this technique could identify embryos that were balanced for the involved chromosomes, it was subject to the same technical limitations described above for aneuploidy testing utilizing FISH. In addition, no information was provided for chromosomes not involved in the translocation, which could produce aneuploidy and decreased reproductive potential. Fiorentino et al.  developed a PCR-based method to evaluate translocations using STR markers that flanked the translocation breakpoints along with others to evaluate the copy number of other chromosomes. This was applied in 27 PGD cycles where 18 couples achieved a clinical pregnancy. Shortly thereafter, mCGH or aCGH was applied after WGA, and this analysis allowed not only aneuploidy evaluation of the translocated chromosomes, but also detected age-related aneuploidy in other chromosomes not involved in the translocation. In the report of Alfarawati et al. , where 16 couples with translocations underwent 20 cycles of IVF/PGD, 22% of embryos were chromosomally normal and 28.9% were balanced for the translocation, but had aneuploidy of other chromosomes. Fiorentino et al.  reported on 28 PGD cycles for translocation in 24 couples with aCGH and found that 16% of embryos were normal for all chromosomes, while 27.3% were normal for the translocation, but showed aneuploidy in other chromosomes. Treff et al.  validated an SNP array for the detection of translocations and evaluated 19 IVF/PGD cycles from 15 patients carrying a translocation. Of 122 normally developing blastocysts, 50.8% were normal or balanced for the translocation and 32% were euploid in the remaining chromosomes. Overall, including arrested embryos, 15.2% of all embryos produced were chromosomally normal. In 12 cases where the embryo transfer of normal embryos was carried out, a clinical pregnancy rate of 75% was seen. The implication of these results was that by using FISH for the PGD of translocations, approximately half of the embryos that were determined to be “normal” actually contained aneuploidy for other chromosomes not involved in the translocation and, therefore, not detected by the FISH analysis used in their case. Less than one quarter of the embryos produced in these cycles therefore were truly euploid and capable of producing a healthy pregnancy. The reproductive outcome in couples carrying a translocation is likely dependent on the shape of the quadrivalent and subsequent modes of segregation of the specific translocation and the risk of producing a viable abnormal gestation. Scriven et al.  argue that for fertile couples carrying translocations with a low risk of producing viable abnormal gestations, natural conception may be capable of producing a normal child in a shorter time and with lower cost than the use of IVF/PGD.