EXPLAINER

Do IVF add-ons work? Two large trials and a regulator urge caution

Popular extras like embryo genetic testing and endometrial scratching did not raise live-birth rates in their biggest randomised trials, and the UK regulator warns most patients need none of them.

The optional extras sold alongside IVF — genetic testing of embryos, endometrial "scratching", immune treatments, embryo glue and more — mostly lack good evidence that they increase the chance of a baby, and the two most popular ones failed to raise live-birth rates in their largest randomised trials [s1][s2][s3]. The UK fertility regulator's blunt summary is that for most patients, routine cycles of proven treatment work without any add-on at all [s1].

What an "add-on" is

The Human Fertilisation and Embryology Authority (HFEA) defines treatment add-ons as optional, non-essential treatments or tests offered in addition to established care such as IVF or intrauterine insemination [s1]. The category it publishes on covers add-ons promoted to the general patient population in UK-licensed clinics, backed by studies that claim to improve the chance of a baby, but where evidence of effectiveness in a clinical setting "is lacking or absent" [s1].

The regulator's consumer warning is unusually direct. If you are paying for your own treatment, it says, you may want to weigh whether the money is better spent on additional cycles of IVF or IUI than on unproven add-ons in a single cycle [s1]. That framing matters because add-ons are typically an out-of-pocket upsell layered onto an already expensive cycle.

The evidence on the two most common add-ons

Two add-ons have been tested in large, well-designed randomised trials, and both came back negative.

Preimplantation genetic testing for aneuploidy (PGT-A) screens embryos for chromosomal abnormalities before transfer, on the theory that selecting "euploid" embryos raises live-birth rates. A multicentre randomised trial assigned 1,212 women with three or more good-quality blastocysts to IVF with PGT-A or to conventional IVF, 606 in each group [s2]. Live births occurred in 77.2% of the PGT-A group, and conventional IVF proved noninferior, with a between-group difference of −4.6 percentage points (95% CI −9.2 to −0.0) [s2]. In other words, adding genetic testing did not improve the cumulative chance of a baby in this population — if anything the point estimate favoured plain IVF. Cumulative clinical pregnancy loss was 8.7% with PGT-A and 12.6% without [s2].

Endometrial scratching — deliberately injuring the womb lining with a pipelle biopsy before a cycle, on the theory that it improves embryo implantation — was tested in a trial that randomised 1,364 women to a scratch or no intervention [s3]. Live birth occurred in 26.1% of the endometrial-scratch group (180 of 690) and 26.1% of the control group (176 of 674) — an adjusted odds ratio of 1.00 (95% CI 0.78 to 1.27) [s3]. There were no significant differences in ongoing pregnancy, clinical pregnancy, multiple pregnancy, ectopic pregnancy or miscarriage, and the procedure caused pain, with a median score of 3.5 out of 10 (interquartile range 1.9 to 6.0) [s3].

How to read a null result here

Neither trial says the add-on harms the average patient's odds; they say it does not help. That is the finding an oversold market is least equipped to deliver, because "no benefit" is not a sellable headline. It is also why the HFEA built a rating system rather than a ban: some add-ons may be appropriate for particular patients or in research, even where population-level evidence of a live-birth benefit is missing [s1].

Two caveats keep this honest. The PGT-A trial studied women who already had three or more good-quality blastocysts, a relatively good-prognosis group, so it does not settle whether testing helps in every scenario clinics market it for [s2]. And a null result in one large trial is not proof of no effect anywhere — but when the best-powered trial of a widely sold add-on lands on the line of no difference, the burden of proof sits with the seller, not the patient. Health Newspapers has examined the same evidence-versus-marketing gap in egg freezing and ovarian-reserve testing.

What it means for a reader

The evidence supports a default of skepticism: proven IVF works for most patients without add-ons, and the two extras with the strongest trials did not raise the live-birth rate [s1][s2][s3]. A reasonable question for any offered add-on is the one the regulator implies — what randomised evidence shows it improves the chance of a baby, and what it costs relative to another full cycle [s1]. This is informational and not medical advice; individual treatment decisions belong with a fertility specialist.

What to watch

Whether clinics' use of add-ons falls as the negative trials accumulate, and whether regulators outside the UK adopt comparable ratings so patients can see, before they pay, which extras have evidence behind them and which do not.

This article is informational and is not medical advice.

Sources

Sources

  1. Treatment add-ons with limited evidence — Human Fertilisation and Embryology Authority , November 1, 2024
  2. Live Birth with or without Preimplantation Genetic Testing for Aneuploidy — New England Journal of Medicine , November 24, 2021
  3. A Randomized Trial of Endometrial Scratching before In Vitro Fertilization — New England Journal of Medicine , January 23, 2019
Related coverage