When Mats Brannstrom of Sahlgrenska University in Stockholm addressed the European Conference on Reproductive Medicine in Madrid three years ago he had a great story to tell. His team had successfully transplanted wombs between female mice and the recipients had given birth to healthy pups. He hoped to try the procedure on humans within three years.

At the same meeting this year Brannstrom said he regretted his remarks in Madrid. While he believes that human uterine transplants will eventually be possible, he has yet to make significant progress in large animals. His biggest success has been to remove a ewe’s womb and put it back into the same animal. The pace of advance will be slower than he first thought.

This is worth remembering amid this week’s reports that Richard Smith of Hammersmith Hospital is two years away from transplanting a human womb. Bold predictions about experimental medical technologies are always best handled with care. They often suit both journalists and scientists, who can enter into a kind of tacit conspiracy; swift timetables for human trials make good headlines and catch the attention of research funders. But they are regularly shown up as guesswork later.

Quite apart from the history of womb transplants, there are grounds to think the prognosis for Smith’s research overambitious. His plan is to take a womb from a dead donor and transplant it into a volunteer who was born without a uterus or who has had a hysterectomy. His animal work suggests the new organ will survive only if attached together with large blood vessels, which would preclude a living donor such as a mother or a sister. Immunosuppressant drugs would be needed to prevent rejection for two to three years, while she attempts to start a family, and the organ would then be removed.

While it is not implausible that a working womb could be transplanted in this way, a big problem remains that threatens to make the project an ethical non-starter. Though organs more complex are routinely transplanted, plumbing a womb into a new body is not the same as attaching a heart or liver. That is because the womb’s blood vessels must not just supply the organ with nutrients, but also a developing foetus. A poor or interrupted blood supply could result in stillbirth, miscarriage, malformations or low birth weight.

This places uterine transplants on a different ethical plane from superficially similar kinds of surgery. A patient receiving a new heart will be dangerously ill to begin with and the hazards of the operation are borne only by herself.

If the recipient of a new womb becomes pregnant, any child she conceives will also be exposed to significant risk. Even milder complications of a poor uterine blood supply could have lifelong effects: there is now overwhelming evidence that babies born underweight have an elevated risk of adult medical problems such as diabetes.

The case for caution becomes stronger still when one considers that there is already a good therapeutic option for most women unfortunate enough to lack their own womb.

Surrogacy is not perfect — most women would clearly prefer to carry their own child, some have religious objections and there have been a handful of custody disputes. But detailed research has shown that with proper counselling it works well for infertile couples, surrogates and their children. Its risks are tiny compared with those of experimental womb transplants that save no lives and could compromise foetal development.

Such operations may ultimately be proved safe and stand up to ethical scrutiny, though that will require much more animal work, including primate studies. But it is wholly unrealistic to expect it all to happen in the next two years.

Mark Henderson is Science Editor of The Times