All over Europe, fertility doctors are under growing pressure to transfer IVF embryos to their patients' wombs one at a time. Standard practice in the UK is to use two, to improve a woman's chances of getting pregnant with at least one, but this has a major consequence: one in four IVF pregnancies produces twins, compared with one in 80 natural conceptions.

Such multiple pregnancies have a higher risk of complications for babies and mothers, including cerebral palsy, prematurity, low birth weight and pre-eclampsia,a life-threatening blood pressure disorder.

This month doctors and regulators launched a strategy to reduce the IVF twin rate to one in ten, which will require half of all patients to have a single embryo transfer. This is common in Scandinavia and the Low Countries, and the European Society of Human Reproduction and Embryology has long been seeking to promote it elsewhere.

Just as its message has been taken up by Britain, however, it came under heavy fire this week at the society's annual conference in Barcelona. Norbert Gleicher, an American IVF specialist, argued that the risks of twin pregnancies have been exaggerated and that many couples should be encouraged to have them. As most infertile couples want at least two children, he said that the risks of IVF twins should be compared with those of two singleton pregnancies, not one. And he presented statistics suggesting that when this is done, most of the hazards disappear.

Gleicher's views certainly chime with those of many IVF patients. Twins would complete their family at a stroke, removing the need for further cycles of expensive and gruelling treatment. But while his analysis was the talk of the conference, it is undermined by several flaws.

First, his suggestion that the risks of singleton pregnancies should be added together and judged against twins is too straightforward. Where issues such as stillbirth are concerned, successive singleton pregnancies are independent events: the chances that both children will be affected are very small. When a twin pregnancy goes wrong, there is a much greater probability that both infants will suffer. What is more, if a mother has a smooth singleton pregnancy first time around, the chances of complications in a later pregnancy are greatly reduced. It is over-simple to combine risks this way.

He also argues that IVF twin risks should be adjusted downwards by 40 per cent, because of data suggesting that IVF twins are often healthier than twins conceived spontaneously. This figure, however, was drawn from a single research paper, and applies to a single risk - neonatal death.

It has no bearing on other dangers, such as pre-eclampsia, yet Gleicher has adjusted his risk ratios by this amount across the board.

Even when risks are revised in this questionable fashion, many remain high: the chances of low birth weight and prematurity are five times greater for twins than singletons. Plus, low birth weight - which affects about 50 per cent of twins - is associated with health problems throughout life, such as obesity and type 2 diabetes.

Gleicher admits that he did not consider this. He also took no account of psychological problems among the parents of twins, including depression, anxiety and sleep disorders.

The medical orthodoxy on multiple births may sound unattractive, but it is probably correct. Where Gleicher has a point, though, is that despite these potential problems, most IVF twin pregnancies produce healthy children. While doctors are right to explain the risks, and to take steps to reduce them, they must be wary of implying that they are greater than they really are.

It is always going to be taxing to carry and care for two children born at once. Parents do not need an added burden of anxiety.

Mark Henderson is the Science Editor of The Times

Explore Health

• Expert Advice

• Health Features

• Mental Health

• Alternative Medicine

• Child Health

• Health Club