Should You Pay for PGT-A? What NICE, ASRM and ESHRE Say

On current guidance, no — not as a routine add-on, and not to improve your chance of a baby. Three bodies say so in writing, most bluntly NICE: guideline NG257, recommendation 1.48.1 (2026), says "do not offer pre-implantation genetic testing for aneuploidy (PGT-A) as part of fertility treatment to improve live birth rates". ESHRE's 2023 add-ons guideline says PGT-A "is currently not recommended for routine clinical use". The ASRM 2024 committee opinion says routine blastocyst biopsy with aneuploidy testing in all infertile patients undergoing IVF "cannot be recommended".

Which test you are being offered matters, because the word "PGT" covers three different things and only one of them is contested:

• PGT-M, for a single-gene condition one or both of you is already known to carry. Established use.
• PGT-SR, for a chromosomal rearrangement such as a translocation already found on a karyotype. Established use.
• PGT-A, which counts chromosomes in every embryo looking for aneuploidy nobody knew about. This is the contested one, and what "PGT" usually means on a price list.

PGT-A does not create more embryos. It sorts the ones you already have, and a cycle that produced one blastocyst has nothing to sort. For the expected proportion of chromosomally normal embryos at your age, our euploid embryo probability calculator sets it out.

Because they are measured on a group the test itself has already filtered. PGT-A is sold on live birth per transfer. What you are buying is a live birth per cycle you started. Those are different denominators, and the gap between them is where the sales pitch lives.

Once the test removes embryos, some patients have nothing left to transfer. They drop out of the per-transfer figure and stop diluting it, and everyone who remains has, by construction, a normal-looking embryo.

The same SART CORS dataset shows both faces of this. Across 133,494 first autologous stimulation cycles, when everyone with blastocysts available for transfer or testing was counted — including the patients left with none to transfer — cumulative live birth was lower with PGT-A at every age band except over 40. When the count was narrowed to PGT-A patients who did reach a frozen transfer, cumulative live birth ran from 71.2% under 35 down to 50.2% over 42. Same cycles, two very different-looking results.

The randomised evidence lands in the same place. Yan and colleagues (NEJM 2021) randomised 1,212 women aged 20 to 37 with three or more good-quality blastocysts, measuring cumulative live birth after up to three transfers within a year. It was 77.2% (468 of 606) with PGT-A and 81.8% (496 of 606) with conventional IVF — a difference of minus 4.6 percentage points (95% CI minus 9.2 to minus 0.0). Cumulative clinical pregnancy loss was 8.7% versus 12.6%. The trial's conclusion was that conventional IVF was non-inferior.

A Cochrane review of 13 trials in 2,794 women, graded low to moderate quality, found insufficient good-quality evidence of a difference in cumulative live birth rate, live birth after the first transfer, or miscarriage. If you are being shown any percentage at all, our guide to reading the success rates you are quoted covers what to ask about the denominator.

There is a subgroup finding, and it is worth knowing exactly how thin it is. The STAR trial randomised 661 women aged 25 to 40 (mean age 33.7) who each had at least two biopsiable blastocysts. Ongoing pregnancy at 20 weeks was 50% (137 of 274) per transfer with PGT-A against 46% (143 of 313) without — and 41.8% (138 of 330) against 43.5% (144 of 331) per intention to treat. In a post hoc look at the women aged 35 to 40, ongoing pregnancy per transfer was 51% (62 of 122) versus 37% (54 of 145), which was significant. Per intention to treat, it was not.

That is the whole case for an older-age benefit: a subgroup, analysed after the fact, significant on the flattering denominator and not on the honest one. ESHRE says as much — the supposition that PGT-A reduces miscarriage or time to pregnancy in groups such as advanced maternal age "is based on post hoc analyses and requires further investigation".

Registry data points the same way at the young end. In 31,900 SART CORS patients aged 37 or under, cumulative live birth among women under 35 was 70.6% with PGT-A and 71.1% without, adjusted odds ratio 0.82 (95% CI 0.72 to 0.93) — worse with the test. Among women aged 35 to 37 there was no association (66.6% versus 62.5%, adjusted OR 0.92, 95% CI 0.83 to 1.01).

Speed is sometimes offered instead: a viable embryo found sooner, even if the end result is the same. The two datasets disagree. In the SART CORS group aged 37 or under, time to a pregnancy resulting in a live birth averaged 4.58 months (SD 3.53) with PGT-A against 2.37 months (SD 3.20) with untested transfers — slower, not faster. ASRM cites one randomised trial in women aged 38 to 41 where it was faster. Ask which group you are in.

They may be discarded, and some of them would have become babies. ESHRE lists "disposal of viable embryos" among the harms of PGT-A.

The clearest evidence is a double-blinded prospective non-selection trial (Capalbo and colleagues, 2021), in which embryos were transferred without the clinic knowing the result. Live-birth and miscarriage rates were equivalent across 484 euploid, 282 low-grade mosaic and 131 medium-grade mosaic embryos, and no mosaicism or uniparental disomy was found in the resulting pregnancies or newborns. On that evidence, a low-to-medium mosaic embryo behaves like a normal one.

Fully abnormal calls are less settled. In a prospective cohort at one New York centre, 69 couples moved 444 PGT-A-abnormal embryos that their original clinics had refused to transfer; 50 of those patients went on to 57 transfer cycles, producing 8 live births and 11 miscarriages with no terminations. One child was born with a segmental duplication and needed repair of a coarctation of the aorta. Mean age at retrieval was 41.4 years, these couples had no other option, and the paper drew three published comments. An abnormal call is not always a non-viable embryo. That is not the same as transferring abnormal embryos being a good default.

ASRM recommends that clinics hold a written policy on the disposition of abnormal embryos and that genetic counselling is readily accessible at any point in the decision. Ask for both before the biopsy. If a transfer has already failed, our pages on recurrent implantation failure and what is worth investigating after a failed cycle cover what the evidence does support looking at.

The Assisted Reproductive Technology (Regulation) Act, 2021 sets a purpose limit. Section 25(1) reads: "The Pre-implantation Genetic testing shall be used to screen the human embryo for known, pre-existing, heritable or genetic diseases only." Whether routine aneuploidy screening falls inside that purpose limit is a legal question this page does not answer, and it is a fair thing to raise with the clinic in writing.

The Act is also explicit about one thing a clinic may not tell you. Section 26(3) provides that no person "shall knowingly provide, prescribe or administer anything that shall ensure or increase the probability that an embryo shall be of a particular sex, or that shall identify the sex of an in-vitro embryo, except to diagnose, prevent or treat a sex-linked disorder or disease". Section 32 makes advertising such a service punishable by five to ten years' imprisonment and a fine of ten to twenty-five lakh rupees. Section 45 states the Act operates in addition to the Pre-conception and Pre-natal Diagnostic Techniques Act, 1994, not in place of it. If any clinic offers you this, it is committing an offence.

One question does most of the work: what would a result change? If the answer is "we would transfer a different embryo first", ask what happens when the test leaves you none to transfer, and what the fee covers then. If the answer is "nothing, but it is good to know", you are paying for information, not a better outcome, and it is reasonable to decline.

• How many blastocysts do I need before selection is worth anything? With one or two, there is almost nothing to sort.
• What is your written policy on embryos called abnormal or mosaic, and will you transfer a low-grade mosaic if it is my only one?
• Is the fee per cycle, per embryo or per batch, and does it cover a re-biopsy if a result comes back undiagnosed?
• Which per-cycle-start number are you quoting me, and is it your clinic's own or from a paper?

None of this applies the same way to PGT-M or PGT-SR for a condition already identified in one of you. Those are targeted tests for a known problem, and the guidelines do not question them. For the rest of the price list: our review of IVF add-ons and what the randomised trials found and of time-lapse embryo selection covers the others, and IVF cost and what Indian law says covers what you can be charged for.

If you have already paid for PGT-A, none of this means the hope was misplaced. The marketing ran ahead of the evidence, and catching that was never your job.

PGT-A is the most heavily marketed add-on in IVF, and the claims run ahead of the trials in specific ways.

• That it improves live birth as a routine test. NICE says do not offer it for that purpose; ASRM says its value as routine screening has not been demonstrated.
• That a euploid result means a pregnancy, or that an abnormal result means a non-viable embryo. Neither follows.
• That it reduces miscarriage. ASRM calls that value unclear; Cochrane found insufficient good-quality evidence of a difference.
• That it gets you pregnant faster. The two largest datasets point in opposite directions by age.
• That it helps because the diagnosis is male factor. ASRM states PGT-A should not be used for that purpose alone.
• That it improves egg quality or yields more embryos. It changes neither.

ASRM's position on consent is the test of whether you are in a real conversation: it should include "a thorough discussion of risks, benefits and limitations of the technology used". If only the benefits have come up, it is not that conversation. Sometimes the answer is not another add-on but more testing, a different protocol, or working with what age does to egg quality.