A screening technique used to select the best embryos during fertility treatment does not provide any benefit for patients, according to three new scientific studies. The technique, which is recommended for some women by the UK's fertility treatment licensing body, may even result in a lowered pregnancy rate.

"I don't think this test is very sensitive or very accurate," said Mohamed Taranissi, who carried out one of the studies and runs the Assisted Reproduction and Gynaecology Centre, a London fertility clinic. "I think people need to understand that before they make their decisions."

He had been a vocal proponent of the technique - called pre-implantation genetic screening (PGS) - when in 2001 the Human Embryology and Fertilisation Authority deliberated over whether to license it in the UK. "I fought the HFEA over it to get it introduced, [but] I'm now looking at it quite sceptically," he said.

PGS involves testing embryos produced during in-vitro fertilisation for abnormal chromosomes. Only the healthy embryos are implanted back into the patient's uterus. "This is a sporadic genetic defect. It can happen to anyone. It is not inherited," said Peter Braude, a fertility expert at Guy's King's and St Thomas' School of Medicine in London. The HFEA recommends PGS for women undergoing IVF who are over 35, have suffered repeated IVF failure, or have had recurrent miscarriages. An IVF attempt can cost £5,000 to £10,000 but PGS adds upwards of £2,000. Seven out of the 85 UK clinics carrying out IVF are licensed to do PGS, used in 50 to 100 IVF attempts in Britain each year. Some doctors have called for PGS to be used much more widely and offered to younger women. A study released in October found that 42% of eggs from all women had genetic defects that could prevent embryos being carried to term.

Dr Jeffrey Nelson, of the Huntingdon Reproductive Centre in California, who carried out the work, said the level of genetic problems in the eggs of younger women was a great surprise. "Just the fact that we are seeing this high rate of abnormality suggests that we should be using [it] more," he added. But although everyone agrees it should work in theory, in practice it is a different story. Professor Marcelle Cedars and Dr Lora Shahine at the University of California at San Francisco have reviewed 12 studies which compared patients undergoing normal IVF with those undergoing IVF with PGS.

Although some studies hint at a benefit from PGS they say the most rigorous ones found no difference between the two. Their research is published in the journal Fertility and Sterility.

"I don't think we yet know why the expected benefit has not come to fruition," Prof Cedars said. "Are we doing some damage to the embryo? Are we testing the wrong things? Are we testing the wrong people?"

One possibility is that although the technique is picking the best embryos, the extra handling and manipulation of the embryo outside the body cancels out that benefit. PGS involves removing one or two cells at the eight-cell stage. The long-term effects on the baby are not known.

Alternatively, an embryo might be capable of fixing its own genetic defects. "What we don't know is the capacity of the embryo to reorganise itself so that it excludes the bad cells and goes on with the good cells," Professor Braude said. "There's good evidence that is the case."

So an apparently faulty embryo might go on to repair itself. Another problem is that the cell analysed may not be typical of the rest of the embryo. "[It] is based on the assumption that all the cells in the embryo are genetically the same, but they may not be," Professor Taranissi said. So a chromosome abnormality restricted to the one cell could condemn an otherwise healthy embryo if that is the one the doctor happens to test. A related technique called pre-implantation genetic diagnosis tests for specific genetic mutations that run in families. There is no suggestion these tests are not effective.

The California study is backed up by another review released this month in the Cochrane Database of Systematic Reviews. Dr Sjoerd Repping of the Academic Medical Centre in Amsterdam, who led the study, said: "There's absolutely no evidence that PGS is more effective." He is finishing a 2½-year study of 400 people into the effectiveness of PGS. The study, which will report in a few months, involved placing patients into study groups at random - the gold standard methodology for medical trials.

More worryingly, there is evidence that PGS might actually make matters worse. Prof Taranissi has compared 79 couples who opted for PGS with 69 who were offered it but declined. His study, not yet published, found that those using PGS were actually less successful than the couples who underwent standard IVF.

James Healy, a spokesman for the HFEA, said the fertility watchdog was confident that PGS was beneficial for the groups of patients it recommended should use it. But he said it would consider the research.

Footnotes

Pre-implantation genetic screening

One or two cells are extracted from the embryo at day three. These are tested for abnormalities called chromosome rearrangements, which occur when part of one chromosome becomes detached and joins to another. These changes - one of which causes Down's syndrome - occur in all women but become more common in older women. PGS, which is also called aneuploidy screening, involves adding DNA probes which stick to common rearrangements. If a probe sticks it glows, denoting a positive result. Only negative embryos are implanted.

Chromosomes

Bundles of DNA and protein that carry our genetic recipe book. Humans have 23 pairs. Our genes are coded in the sequence of DNA letters.

In-vitro fertilisation

The standard fertility treatment involves fertilising extracted eggs outside the body then reimplanting healthy looking embryos. The embryos can be kept out of the body without freezing for up to five days.

Pre-implantation genetic diagnosis

Essentially the same as PGS, but it tests for specific genetic diseases that run in families. Only couples who have a family history of the disease (for example Huntingdon's disease) need undergo the test.