The ethics of preimplantation genetic testing using polygenic scores

The increasing accessibility of genomic data, coupled with a general interest in wanting to learn more about our genetic makeup has seen commercial interest in offering preimplantation genetic testing to people undergoing IVF, enabling the testing of embryos for genetic traits before transfer to the uterus. It's important for potential parents, stakeholders and the wider public to understand the credibility of the claims surrounding these technologies, which we attempt to address in this short piece.

What's new?

Preimplantation genetic testing (PGT) is highly regulated in the UK by the Human Fertilisation and Embryology Authority (HFEA) and permissible where there is a particular risk of transmission of a serious monogenic disorder (PGT-M) or structural chromosomal rearrangement (PGT-SR) and where there is knowledge of the specific causative genetic variant(s). PGT is available to NHS patients who have a family history of genetic illness and who satisfy other eligibility criteria; some people may also opt to fund this treatment privately.

PGT can provide reassurance that a high-risk genetic disorder will not be passed onto offspring, however, some private sector companies are now also offering PGT; that is, testing of embryos for genetic markers in the absence of a family history of genetic illness, with the intention of predicting an embryo's future traits, conditions and health risks. On offer are a range of techniques to screen embryos for monogenic childhood- and adult-onset disorders, chromosomal abnormalities and gene carrier status for autosomal recessive disorders, as well as polygenic scores, pharmacogenetic variants, de novo variants (not seen in previous generations of a family), novel variants (not reported before) or rare variants and mitochondrial DNA analysis.

What's the problem?

Advances in genetic testing technologies have made determining the sequence of the genetic code affordable, yet its interpretation, and what it means for health is often much more complex. In a clinical setting, genomic data are looked at alongside signs and symptoms, family history, and other details in order to make conclusions about their connection to health issues. All individuals have rare and de novo variants present in their genome which may have little, if any health implications. Likewise, everyone carries gene variants for autosomal recessive conditions yet this will only have potential clinical consequences for future generations. Testing for both carrier status and adult-onset conditions is discouraged in childhood unless there are interventions to offer.

When organising genetic tests for a suspected rare disorder (such as a single gene or chromosome disorder in which there is a strong genetic component), variants tested for are usually selected are usually selected due to their likelihood of being associated with a genetic condition or group of genetic conditions. Most common disorders are caused by many genetic variations, as well as environmental and lifestyle factors interacting together. Polygenic scores (also known as polygenic risk scores) rely on the sum of predicted associations between multiple, common, genetic differences across the genome and a trait or condition, yet do not capture the complexity of the environmental and lifestyle interactions that often play a greater role in common disorders.

Polygenic scores have particular uses in medical research and understanding population-level risk, but on the whole, many are poor predictors of risk at an individual level. For example, common genetic differences often affect multiple traits, which means that 'positively' viewed traits might be accompanied by other traits which are considered less favourable. What's more, the predictive power of polygenic scores can vary both for an individual and on a population level by how they are calculated, in which populations they were derived, and which populations an individual is compared to.

To further complicate matters, many of these scores are based on data from people who are mostly of either high socioeconomic status, good health, European ancestry or some combination thereof, which pose challenges to their transferability to other populations. Many researchers are working to address these challenges, but there is still a long way to go. These complexities make polygenic scores difficult to interpret in the context of assessing an individual's risk.

Is it a good idea?

While discussions about genetic testing often veer into debates about eugenics and the concept of 'designer babies', fixating on these dystopian scenarios can detract from the less glamorous but essential present-day concerns – that is, the fact that these tests are premised on providing information to aid in decisions about which embryos to transfer, but that much genetic data will not help with that decision.

Preimplantation polygenic tests meet few, if any of the well-established and widely accepted criteria already in use to determine the usefulness of medical screening tests, and in particular, to reduce the risks of false positives and over-investigation. The limitations of preimplantation polygenetic tests are unlikely to be clear to people being offered these options as part of their IVF journey or for those individuals who, while not currently going through IVF, may find the possibility of selecting certain genetic traits for their prospective child attractive.

There is public interest in preimplantation genetic tests, yet the quest to find a genetically 'normal' embryo is likely to lead to disappointment. We all carry genetic variants, the significance of which may or may not become evident during our lifetime. In short, there is no such thing as a 'normal' embryo. Individuals/couples undertaking PGT in the hope of finding a perfect embryo may well find genetic 'abnormalities' in all of their embryos, which they might then discard. Or, they may choose to go ahead with transfer of an embryo with an 'abnormal' result but believe it to be imperfect from conception, causing unnecessary health anxiety and potentially placing unrealistic demands on health services postnatally. Lastly, they may be falsely reassured by 'normal' preimplantation test results only for their child to later be diagnosed with a genetic condition.

Neither preimplantation genetic tests for aneuploidy nor tests using polygenic scores have been shown to increase the chances of a successful pregnancy, and such tests are not funded as part of NHS treatment. Individuals/couples accessing these tests through the private sector need to be informed of the limitations of their use.