Polygenic embryo screening estimates genetic susceptibility to some diseases; it cannot determine a future child’s health. The scores combine many genetic variants into probabilities, while the clinical value of choosing between embryos on that basis remains unproven. The UK’s fertility regulator says this use, commonly called PGT-P, is not lawful in the UK.

The distinction is between estimating risk, demonstrating a benefit from using that estimate and being permitted to use it in a particular healthcare system. Those are three separate questions.

What the score is trying to predict

The American Society for Reproductive Medicine’s 2026 ethics opinion describes PGT-P as an attempt to estimate susceptibility to conditions influenced by many genes and other factors. It involves in vitro fertilisation (IVF) and genetic analysis of embryos, with the aim of using scores to inform selection.

Unlike a test for a particular inherited gene variant, a polygenic score combines small contributions from many locations in the genome. It is probabilistic: a lower estimated risk does not mean a child cannot develop the condition, and a higher score does not establish that they will.

ASRM identifies limits in applying models largely developed from adult populations of European ancestry to different populations and to children whose environment and healthcare may be very different decades later. The opinion does not recommend PGT-P as a clinical service at this time.

PGT-P, PGT-M and PGT-A ask different questions

The shared “pre-implantation genetic testing” name can obscure important differences:

  • PGT-M checks for a specific inherited genetic condition. The HFEA’s patient information explains its distinct permitted-use framework in the UK.
  • PGT-A screens for chromosome-number differences, known as aneuploidy. The HFEA’s PGT-A information addresses that separate test and its evidence; it does not validate polygenic risk selection.
  • PGT-P estimates susceptibility to multifactorial conditions using a collection of variants. Evidence and permission for another form of embryo testing cannot simply be transferred to it.

These are definitions, not a ranking of which test a particular person should have.

A risk reduction needs a starting risk

ASRM also distinguishes absolute risk from relative risk. A percentage reduction can sound substantial while describing a small difference in the probability of developing a condition.

For illustration only, a fall in estimated risk from 2% to 1% is a 50% relative reduction but a one-percentage-point absolute reduction. These are invented arithmetic inputs, not measured PGT-P outcomes or a prediction for an embryo. Neither description would establish that selecting an embryo on the score actually produces the projected health benefit.

There is a second practical constraint: an IVF cycle yields a limited set of embryos, not an unrestricted menu of genetic profiles. Ranking within that set must be distinguished from comparing unrelated adults in the population used to develop a score. ASRM notes that some cycles produce too few embryos for ranking to have a meaningful effect.

What the UK position means

The Human Fertilisation and Embryology Authority’s current explanation says PGT-P does not meet the permitted purposes for embryo testing under UK law. It also says UK embryo selection cannot use prohibited information simply because the analysis was performed overseas.

The HFEA warns that excluding embryos on an unproven risk estimate may reduce the number available for transfer, including embryos that would not have developed the predicted disease. This is a potential consequence identified by the regulator, not a measured reduction in live births that can be assigned to every patient.

The UK legal position and the US professional opinion have different roles. Marketing or availability in another country is neither UK permission nor evidence of long-term benefit.

What evidence would change the question?

A useful evaluation would need to show how the score performs for the intended population and whether acting on it improves outcomes that matter, while accounting for the limited embryo set and possible tradeoffs. Retrospective risk modelling is not the same as following children after a selection strategy.

For now, a score can describe an estimated susceptibility under a model. It cannot supply a guarantee of health, a complete comparison of every future disease risk or an evidence-based instruction about which embryo an individual family should choose.